MediaWiki talk:Monobook.css

Gadget

There is now a "gadget" available through user preferences for showing rounded corners. --JWS 20:15, 29 December 2007 (UTC)

2006 discussion

Please remove the rounded corners. It looks awful. Browsers cannot render it appropriately. They're all jaggy and aliased. Besides, not a single other English Wikimedia project uses this. It's the box object equivalent of <blink>. 85.147.58.212 21:50, 20 August 2006 (UTC)

1. I'm not going to.
2. No point in arguing over personal preference.
3. Appropriate is in the eye of the beholder.
4. They don't appear like that on my screen. Nobody from the project has complained.
5. That nobody else is doing something is not an argument unless you actuall show why that is.
6. I'm not going to dignify the last comment with a response.

-- sebmol ^? 21:59, 20 August 2006 (UTC)

"Rounded" corners

I've been away for a while, but I see that you once again enabled the "rounded" corners in November. Was this discussed somewhere? Not only does the anti-aliasing issue (which makes the effect look terrible for many users) remain, but such styling clashes with the main page design (which actually was determined via community consensus). And of course, it's nonstandard code that doesn't work for most people. —David Levy 14:29, 8 January 2007 (UTC)

---

In reply to David Levy from Robert Elliott

So far, I have found rounded corners to work well. Rounded corners works with Fox Fire but not with Safari. With Safari, rounded corners revert to straight lines which looks fine. In both cases, there is no problem.

Currently, "rounded corners" is only a small part of the effort to make the pages of Wikiversity more attractive to students. Rounded corners work extremely well for students because they make the page easier to view. Look at Learning the Basics of French to see how pages are improved with rounded corners.

In the past, all of Wikiversity's efforts have been to attract people who will write lessons (which are vital in getting Wikiversity started.)

Now, Wikiversity must begin to attract people who have no interest in writing lessons but rather, want to learn something. These are the typical students who will use Wikiversity.

As has been pointed out recently, the current MAIN PAGE is designed to attract people who want to participate in building Wikiversity but not people who simply want to use Wikiversity. Therefore, the current main page emphasizes the structure of Wikiversity but does not emphasize fully working lessons. The current main page points mostly to stubs and empty pages which is an immediate turn-off to potential students.

For that reason, the main page must eventually be replaced by a more consumer-friendly page designed to attract students who have no desire to be authors of lessons but simply want to learn new and exciting things. Robert Elliott 15:38, 8 January 2007 (UTC)

Did you view the image that I provided? That's an actual screenshot (magnified 5x) of how the rounded corners look in Firefox for Windows (the most widely used browser that supports this feature). Here is 1x version.

As you can see, they appear not as rounded, but as a series of jagged, partially disconnected lines. This obviously harms the site's appearance and does nothing to attract students or anyone else. I assume that the effect looks much better in OS X, but that operating system is used by a small minority of people.

Furthermore, most people use Internet Explorer, which does not support this feature. How are we to create a uniform design when we deliberately code the pages in manner that causes different people to see different things? We should strive for a standardized appearance across as many graphical browsers as possible. —David Levy 16:30, 8 January 2007 (UTC)

Interesting notion but in my opinion rather incompatible with the way the web works. There's no guarantee sites look the same across browsers or platforms. If rounded corners appear in Firefox but not in IE, it's a shame. But it's not something I'm going to be awfully concerned about because it's a purely visual issue that does not hinder anybody's abilities to use the site. If you have such a huge problem with them, I suggest adjusting your monobook. If you don't think that solution is sufficient and the problem is of a significant priority for the community, feel free to raise the question at the Wikiversity:Colloquium and start a broad discussion about it. sebmol ^? 15:24, 9 January 2007 (UTC)

1. There's a major distinction between minor, unavoidable differences and major style differences deliberately introduced to code. The corners are the only element preventing the site from looking virtually identical across all of the major browsers and operating systems.

2. The reason why IE (and most other browsers) don't support this feature is that it is not included in the HTML standard. It is a very bad idea for a public, open-source project to use non-standards-compliant HTML code.

3. I could just as easily suggest that you adjust your personal MonoBook code. We both know that very few people do this (or even know how to), so the settings applied to the site's CSS determine what most users will see.

4. I (and others) did raise this issue at the Colloquium, and you initially dismissed our opinions after deciding that we weren't members of the community. I pointed out that this is an aesthetic matter (not an editorial one), and you eventually relented. Then you waited a while and (knowing that this was controversial) reinserted the code without any discussion that I'm aware of. (As far as I can tell, you simply like the way that this looks.) Do you believe that your position as a sysop grants you the authority to unilaterally make these decisions on behalf of Wikiversity (while demanding that non-sysops "raise the question" "and start a broad discussion about it")? Given the fact that straight corners are the MediaWiki default, I believe that it's incumbent on you to establish consensus for the rounded corners. At the English Wikipedia (where I'm a sysop), I hold myself and others to the same standard.

5. Why are you ignoring the fact that the corners look like this for many people? You previously acknowledged that you weren't aware of this when you originally added them, but now you are. As you correctly noted above, this is "a purely visual issue," so I don't understand how you can justify making the site look terrible on many people's machines. —David Levy 16:39, 9 January 2007 (UTC)

I introduced the rounded corners fairly early at the start of the project. You complained about them together with a few others claiming things which IMO have rather little relevance to Wikiversity. For all intents and purposes, at this project, rounded corners were the default. It really doesn't matter though: There are Wikimedia projects which have rounded corners and those that don't.

As you know, this isn't the English Wikipedia nor do we strive to be, act or look like it. One major advantage is that we lack their stifling excesses in process and take WV:BOLD much more seriously here. That means, when something needs to be done, we don't spend endless time discussing it, building consensus and making everyone happy. We just do it. Otherwise we'd still be debating on how to name our categories or what layout topic pages should follow or what the appropriate process for speedy deletion should be, all that without actually producing anything.

I've given you options to bring this matter of yours to broader attention. The last discussion happened on the talk page of Wikiversity:Main Page, not Wikiversity:Colloquium, our most active general discussion page. So go there and raise your question. If there are other custodians who agree with your arguments, they'll surely undo my change. sebmol ^? 18:31, 9 January 2007 (UTC)

I introduced the rounded corners fairly early at the start of the project.>

Yes, and several users complained almost immediately.

You complained about them together with a few others claiming things which IMO have rather little relevance to Wikiversity.>

Please explain how the appearance of Wikiversity has "little relevance to Wikiversity."

For all intents and purposes, at this project, rounded corners were the default.>

No. This is a decision that you made unilaterally and that was criticized by numerous people. In no way is it sacrosanct.

Because of the controversy, you eventually removed the code in question. Without any attempt at further discussion, you then unilaterally restored it. Now you claim that it's my responsibility to initiate discussion and establish consensus against your unilateral, HTML standards-defying change. Why? Because you're a sysop and I'm just a regular user.

It really doesn't matter though:>

It doesn't matter that the site looks like this to many people?

There are Wikimedia projects which have rounded corners and those that don't.>

You know perfectly well that none of the other English-language Wikimedia sites use the rounded corners. I don't speak any other languages, so I'm unable to raise the issue at the non-English sites.

As you know, this isn't the English Wikipedia nor do we strive to be, act or look like it. One major advantage is that we lack their stifling excesses in process and take WV:BOLD much more seriously here.>

"Boldness" refers to acting without prior discussion. It does not refer to ignoring concerns after they are raised.

That means, when something needs to be done, we don't spend endless time discussing it, building consensus and making everyone happy. We just do it.>

Firstly, please explain why the introduction of rounded corners "need[ed] to be done."

Secondly, I'm not claiming that you did anything wrong by adding them in the first place. But when it quickly became clear that substantial opposition existed, it was not appropriate to dismiss it. It was even more inappropriate to reinstate the change (which by this point had been discussed and had proven controversial) without any attempt at further discussion. This implies a belief on your part that your opinion supersedes others'.

Boldness is fine. I frequently edit MediaWiki pages at the English Wikipedia without consulting anyone. In most cases, no one complains. When someone does, however, I don't claim that it's their responsibility to prove that my changes were bad. The onus is on me to prove that the changes were good. A sysop bit is a means of working on behalf of the community, not a license to overrule non-sysops.

I've given you options to bring this matter of yours to broader attention.>

Again, I already did that. You're the one who has bypassed this step.

The last discussion happened on the talk page of Wikiversity:Main Page, not Wikiversity:Colloquium, our most active general discussion page.>

To what discussion are you referring? I'm referring to this one (which most definitely was conducted at the Colloquium). You, conversely, have made no attempt to initiate discussion regarding this matter on any page.

So go there and raise your question. If there are other custodians who agree with your arguments, they'll surely undo my change.>

You continue to imply that sysops possess the sole authority to make these decisions. That isn't so.

And once again, why are you ignoring the fact that the rounded corners look like this for many people? Do you believe that it doesn't matter how they look on anyone's screen but yours? —David Levy 19:28, 9 January 2007 (UTC)

David, if you don't like rounded corners, just add the relevant line to your monobook.css and it will display square corners. On the French-language Wikipédia, we have been using rounded corners for years and nobody has ever complained about corners preventing people from browsing happily. Sp please stop arguing about such a detail, and think about how much useful content you could have written in the meanwhile (in French we call such useless arguing "de l'enculage de mouches", whose translation "fucking flies" isn't really exact). guillom 11:00, 11 January 2007 (UTC)

I'm not claiming that this affects people's ability to access information. As Sebmol noted above, this is "a purely visual issue." Numerous people have criticized this element, but Sebmol doesn't care. He likes the rounded corners, so that's that. He previously acknowledged that the effect looks much worse in my screenshots than it does on his computer, but the latter is evidently all that matters to him.

I previously raised this issue at the Colloquium (which Sebmol now denies), and it was demonstrated that the inclusion of the rounded corners was controversial. As a result, Sebmol eventually relented and removed them from the code. Then, without any attempt at further discussion, he simply restored them. Now he claims that it's my responsibility to initiate yet another discussion regarding his non-standards-complaint code.

If I do that, and Sebmol once again is persuaded to remove the code, will he simply add it back when he feels like doing so? Will it then be incumbent on me to "start a broad discussion" yet again (and every time that this strikes Sebmol's fancy)? —David Levy 12:47, 11 January 2007 (UTC)

Are you quite through assuming bad faith and misrepresenting the situation? There were not "numerous" but only two users who complained about the rounded corners: you and an anonymous user, whose sole contributions to this project have been that complaint.

The image you used to demonstrate how awful the corners look doesn't really prove anything because the same effect can be achived by magnifying any part of the interface (I did that earlier by magnifying a portion of text from this edit screen Media:Anti-Aliasing Demonstration.GIF). It's always going to look awful if highly magnified because screen resolution is typically small and always finite.

I also didn't deny that there was a discussion at the Colloquium, I simply forgot about it. I've looked at it again today to see what I didn't remember but that one doesn't really amount to much either other than you complaining about it. sebmol ^? 14:16, 11 January 2007 (UTC)

1. I am not assuming bad faith on your part. I believe that you have committed errors in judgement, but I don't attribute your actions to malice.

2. At least two anonymous users (85.147.58.212 and 83.78.187.95) complained about the effect, but I was referring to numerous users from the English Wikipedia (which is large enough to attract much wider feedback than can be found at a relatively new project). As I noted in my original post, I first learned of this issue my reading comments from numerous Wikipedians regarding how ugly and unprofessional the rounded corners made Wikiversity look. This upset me a great deal.

You, of course, claim that such opinions are irrelevant because these individuals aren't members of the Wikiversity community. I strongly disagree. While Wikipedians aren't necessarily qualified to comment on most Wikiversity-related matters, you noted yourself that this is "a purely visual issue." Even someone who's never edited Wikiversity is qualified to express such an opinion, as it pertains strictly to the configuration of MediaWiki (not to Wikiversity's editorial content). When it comes to matters of visual design, we're all members of the Wikimedia community (with people who aren't even active in certain projects helping to design their logos).

3. I provided two images. This one illustrates the site's appearance at actual size. I included the 5x version only to clarify why the corners look so bad at actual size. (Your example image shows no such imperfections.)

4. I uploaded those images per your request. From the point at which the first report of "jaggy and aliased" corners was posted (on this page), you noted that "they don't appear like that on [your] screen." Upon seeing my images, you acknowledged that "they don't appear as jagged on [your] screen as on the screenshot." Now that you realize that this issue exists for many people (albeit not for you personally), why do you continue to believe that the use of rounded corners is a good idea?

5. When you removed the code "in the interest of civility and constructive discussion," I took this to mean that you would actually initiate "constructive discussion" to determine consensus for the non-standard code before restoring it. Instead, you claim that it's my responsibility to initiate another discussion and establish consensus against the change. —David Levy 01:33, 12 January 2007 (UTC)

"Numerous people have criticized this element" --> You are actually the only one (with the IP above) complaining about that, although I'm pretty sure you are making as much noise as a hundred users. guillom 14:22, 11 January 2007 (UTC)

As noted above, I'm referring to comments made about Wikiversity at the English Wikipedia. Unfortunately, Wikiversity is not yet active enough to attract that level of discussion here. —David Levy 01:33, 12 January 2007 (UTC)

A possible solution to this problem would be to revert Monobook to the square edge style and create another style sheet MonobookRounded. Then people can choose which one they like. Personally I agree with David the rounded corners look horrible. If we must have rounded corners would it be possible to use this elegant solution instead?. Finally shouldn't we be striving to uphold web standards and have cross browser compatability. mystictim 11 January 2007

For some reason, the round corners for the buttons do seem to look worse for Firefox in Windows than in OSX, although I personally long ago gave up worrying about the defects of Windows. The terms "horrible", "terrible" and "clashes" .... do they really apply to the minor jaggies of the rounded corners? I like the rounded corners and would like to keep using them. I do not understand the relevance of the arguments about rounded corners being non-standard. I would prefer that Wikiversity have some distinctive features rather than march in lock step with people who think diversity should be prevented. --JWSchmidt 21:39, 11 January 2007 (UTC)

Is there anyway to improve the rounding effect? Like making the rounding smaller to the point that the toolbar looks strait and the bigger boxes look like the rounding done in the toolbar? Or to do the toolbar and the boxes code separate?--Rayc 23:54, 11 January 2007 (UTC)

I tested with my own monobook, and the code that sembol added and subtracted only seems to effect the toolbar. I'm not sure how the other parts are changed.--Rayc 00:12, 12 January 2007 (UTC)

1. While Windows has had more than its share of defects, this isn't one of them. This is a major problem with Mozilla's Windows implementation (which is to be expected, given the fact that this is an experimental feature).

2. While this is entirely subjective, yes, the corners look extremely bad in Firefox for Windows. I've yet to show them to single person who didn't agree. Far more people use the Windows version of Firefox than use the OS X version (because of the relatively small number of OS X users and Safari's superiority to IE), but my stance would be the same if the situation were reversed (and the corners looked good in the Windows version and bad in the OS X version). The straight corners display properly in every major browser, so it would be unreasonable to forsake even a tiny minority of the community.

3. Wikimedia projects set the de facto standard for MediaWiki wikis. As such, it's important that we not encourage bad coding practices. The use of browser-specific, non-W3C-compliant code sets a nasty precedent that can only lead to worse problems. A similar situation arose when Microsoft began introducing nonstandard features to Internet Explorer. What began with relatively minor effects that failed gracefully eventually led to websites that were literally unusable in Netscape Navigator. The W3C standards exist for a reason, and we should strive to follow them.

4. No one is arguing that Wikiversity shouldn't have a distinctive appearance that sets it apart from the other Wikimedia projects. This code actually prevents us from establishing one by creating needless visual disparity among the major browsers and operating systems. Most people don't see the effect at all. Of those who do, many see it only when they use a certain browser, so the site's appearance doesn't even remain consistent for them. It's common, for example, for people to use Firefox at home and IE at school (because Firefox isn't available there). —David Levy 01:33, 12 January 2007 (UTC)

Speaking as a man on the street who just happened to stumble upon this discussion, I have to say that I absolutely agree with David. As I experience them, the unsightly, bulbous tabs which protrude from the top of this and every other Wikiversity page are embarrassingly unprofessional. I only wish you could truly see to what extent they are improperly rendered. They aren't rounded at all, but rather... squashed. I am the archetypal "consumer" you were referring to - a student, using a budget Windows PC and Firefox - and the corners make me cringe. Zioroboco 03:41, 12 January 2007 (UTC)

I'm using a 4 year old "budget" graphics card and the screen shot (below) was 1024x768 screen resolution. I admit that the curves look "drunken" or jagged (as previously discussed), but I do not understand how the term "squashed" applies. --JWSchmidt 04:46, 12 January 2007 (UTC)

The one element that I would describe as "squashed" is the "+" tab (due to the small size and resultant closeness of the corners). —David Levy 05:34, 12 January 2007 (UTC)

I was thinking specifically of the "+" tab when I wrote that. A few other elements which also find themselves lacking comfortable white space and symmetry would include the top right hand corner of the edit tab, the top right hand corner of the search box, and the top left hand corner of the body of the page, where it meets the "article/message/whatever" tab. The more I look, the more examples I find. There's absolutely nothing wrong with sharp edges... I can't believe this has become such an issue. Forgive me if I've gone and poked my nose where it doesn't belong - maybe this is just how things get done, here. Zioroboco 08:44, 12 January 2007 (UTC)

a few notes

I took a look at the corners on a few different platforms. Using Opera on unix the corners are square. This is probably because Opera does not attempt to render non-standard html. With Firefox on a laptop running XP (1024x768 display) they are round, but jagged. With IE on the same XP laptop they are square. It seems that only a few browsers attempt to render the rounded corners, and when they do it often looks bad. Given that Firefox on Windows is a very common platform, my feeling is that the corners should be left square until such time that browsers render them in a consistent way.--mikeu 16:28, 23 January 2007 (UTC)

Hello?

The most recent messages on this page (including the ones by me) appear to have gone ignored. I've been rather patient, and it would be nice if Sebmol, JWSchmidt and Guillom (who seemingly accused me of sock puppetry and also has ignored messages at User talk:Guillom) would respond. —David Levy 01:15, 7 February 2007 (UTC)

Things go slowly around here. I've been trying to find the mediawiki extentions folder for over a month now. :) Did you respond to my comment on coding in the top tabs and the toolbox corners separatly? I think that the tookbox corners look fine, where the edit tabs could be improved by toning down the rounding a bit.--Rayc 02:00, 7 February 2007 (UTC)

I don't know whether those elements can be modified separately, but keep in mind that what looks fine on your computer might not look fine on someone else's. (All of the rounded corners are horribly jagged for me.) Only the straight corners look the same in all major graphical browsers.

It's a bad idea to use non-W3C-compliant code that works for a minority of people, looks terrible on the systems of some of those affected, and renders the idea of maintaining a uniform cross-browser appearance (or even a consistent appearance for specific users) impossible. —David Levy 02:15, 7 February 2007 (UTC)

I am at a loss to contribute anything new to this discussion. I like the rounded box format. A few jaggies in the curves have never bothered me. I fail to understand why this is a matter for continuing discussion. --JWSchmidt 04:35, 7 February 2007 (UTC)

1. No one has addressed my points made at 01:33 on 12 January (UTC).

2. No consensus in favor of this nonstandard software hack has been demonstrated. Why does it remain in place? —David Levy 04:50, 7 February 2007 (UTC)

---

1) "major problem with Mozilla's Windows implementation" <-- I'm not sure I would ever call this a major problem.

2) "the corners look extremely bad" ..... "I've yet to show them to single person who didn't agree" <-- I like the rounded corners; I guess I do not count as a person.

3) "bad coding practices" <-- I like the round corners; I guess I am "bad" (Axis of Evil ID card available upon request). "The W3C standards exist for a reason" <-- I do not think the world will end if Wikiversity has rounded corners.

4) "needless visual disparity" <-- who ever said that humans are logical? I like the rounded corners.

2b) "No consensus in favor" <-- "Silence equals consent" is the ultimate measure of consensus.
--JWSchmidt 23:05, 7 February 2007 (UTC)

1. Everything is relative. This is not a major problem in the grand scheme of things, but it is major as graphical flaws go.

2. I was referring to the people to whom I've shown this in person. I don't know exactly how the corners look on your screen. You described seeing "a few jaggies," but I see nothing but jaggies.

3. I never equated this to any sort of "evil" or claimed that any person was bad.

4. Yes, you like the rounded corners. Other people dislike the rounded corners. By "disparity," I was referring to Wikiversity's highly variable appearance from one browser to the next, not to its differences from the other Wikimedia wikis.

2b. There hasn't been silence. Several people have complained. Numerous others at Wikipedia mocked Wikiversity for its ugly, unprofessional appearance. (I don't agree with this tack, incidentally.) As this is a nonstandard software hack, the onus is on its advocates to demonstrate consensus.

Also keep in mind that only ~12.5% of users even see the rounded corners. The remaining users (~87.5%) are unaware of this element's existence, so their silence cannot be construed as tacit approval.

In any event, thank you for responding. —David Levy 01:12, 8 February 2007 (UTC)

Just as a matter of interest, could you (David) link to that/those discussion(s) on Wikipedia which mocked Wikiversity? Thanks. Cormaggio ^beep 11:44, 8 February 2007 (UTC)

I was only able to find the brief exchange in which I participated (by searching my contribution history). Please take note of my concern regarding the manner in which I was treated here. You've consistently been friendly and courteous (and I sincerely appreciate that), but others (Sebmol and now Guillom) have been downright nasty. Had I been new to Wikimedia, I probably would have been scared away (which seemed to be their goal, given the fact that they told me to go away).

There have been far more scathing (and impolite) assessments of Wikiversity's rounded corners, but I don't recall the talk pages on which they occurred (and I certainly didn't take part). —David Levy 17:27, 8 February 2007 (UTC)

Your edit to MediaWiki:Monobook.css

moved from my talk page where it doesn't belong. guillom 20:12, 7 February 2007 (UTC)

1. Thank you for the effort, but the corners look every bit as jagged as before.
2. Regardless, this doesn't address my other concerns.
3. Are you accusing me of sock puppetry? I assure you that I've posted under this name alone. —David Levy 23:05, 17 January 2007 (UTC)

I would sincerely appreciate a response. —David Levy 07:02, 29 January 2007 (UTC)

Again, I await your response. You've edited the site after both of the above messages were posted, so you must be aware of their existence. I don't appreciate being ignored any more than I appreciate your apparent allegation of sock puppetry. —David Levy 01:15, 7 February 2007 (UTC)

It seems you have much time to loose arguing about details and acting like a capricious child. Good for you. I haven't. Thus, I have chosen to ignore your messages. To my mind you haven't done anything constructive on this project, all you have done and go on doing is wasting everybody's time. Come back to me when you are open-minded, mature and (last but not least) humble. guillom 16:09, 7 February 2007 (UTC)

Firstly, I'm curious as to why you believe that the hours upon hours that I invested in helping to redesign this site's main page don't constitute "anything constructive."

Secondly, is it safe to assume that Wikiversity has no policy against personal attacks? I find it odd that you would respond to my polite inquiries by simultaneously hurling the above insults and question my maturity. I merely asked you to address my concerns and explain why you apparently accused me of sock puppetry. I don't see what's so unreasonable about that. —David Levy 17:09, 7 February 2007 (UTC)

Your POINT is really boring. guillom 20:12, 7 February 2007 (UTC)

Launching a new debate in February 07

I think David makes an excellent point regarding compliance with standards and publishing code that works well, if not perfectly, across the broadest range of browsers because it is standards compliant. Certainly exceptions could easily be made in specific learning projects or sub pages where people want to do their own thing but the Wikiversity main page should standards compliant. Mirwin 03:26, 8 February 2007 (UTC)

"Wikiversity main page should standards compliant" <-- Why? --JWSchmidt 03:44, 8 February 2007 (UTC)

Because it's unprofessional to use experimental, nonstandard HTML code that delivers unpredictable (and often undesirable) results. How many professional websites utilizing this code can you name? Excepting some of the non-English Wikimedia sites, I'm not personally aware of any. —David Levy 04:10, 8 February 2007 (UTC)

Because adherence to specified standards is critical to keeping any complex system functional and useful. The internet is a complex system and it is mission critical as our only operations infrastructure. We should help it work well, not put undue strain on it or its users. We wish people to routinely arrive at the Wikiversity main page be impressed, continue to help files or help desks or portals, be impressed, and work their way to areas of personal interest. If they find an ugly malfunctioning screen in an area of personal interest they are likely to stick around and help fix it. If they find an ugly malfunctioning screen on the main page, test it via on of the standards testing utility sites now available and find it deficient .... what? .... Maybe they stick around maybe not. If they bothered to test and find the utility site confirmed the problem is ours and not their browser's and then mention it in passing on the page design chat page only to find standards do not matter locally... what? Most engineers and computer specialists I know would conclude we have some serious deficiencies in our operational model at Wikiversity if we cannot even get our main page correct. As I get the chance I will ask around and see how many would stick around and help fix it and how many simply write us off as technically incompetent and not worth the time. Incidentally, has anyone run the main page through a test suite to see if the html is correct or not correct? In legalize is it stipulated as fact by all present that we are not currently compliant with html standards set out by the WC3? Mirwin 05:00, 8 February 2007 (UTC)

I am curious. Has anyone passed this non standard HTML by Brion (server administrator last I knew) for an opinion about what it does to the performance of our servers rendering each and every page served from non compliant html? Does the Wikimedia Foundation have any applicable policies regarding compliance or noncompliance with technical standards for things such as HTML, Javascript, etc. for components delivered to the web by the Foundation Servers? Not the information represented or discussed but the underlying technical components, protocols, etc. used to serve (handshake with) the content to web at large. Mirwin 05:54, 8 February 2007 (UTC)

MediaWiki itself creates HTML that's not compliant with W3C standards. As a result, Wikimedia has W3C's Tidy installed on its servers to strip out the most egregious errors. That said, Wikimedia isn't the kind of organization where many things are handed from top down, especially not as trivial issues as what HTML standards to use. This sort of thing clearly falls within the responsibilities of each project. sebmol ^? 06:55, 8 February 2007 (UTC)

Interesting! You are entitled to your opinion, in my view standards compliance is not trivial. Perhaps we should launch a parallel debate on the tech developers mailing list and find out the local pros and cons of compliance and what kind of effort would now be required to get within compliance or even if it is perceived (by the tech wizards) as possible at this point in time. Jimbo could always issue a Jimmy Flashgram should a consensus emerge amongst the developers and/or administrators that they could not sell to the general population. We also might be able to influence the W3C that our methodologies are worth including in the next published/approved version of html standard so the entire internet can benefit from our local genius and innovation. While unlikely, it would obviously make our local debate regarding standards moot. Any idea what percentage of CPU time of the operational servers is now expended on running Tidy between Wikimedia and the Browsers? Any idea whether Tidy imposes its own constraints of potential interest to Wikiversity Departments such as ... no SVG, or no JavaScript, or no Tex inline or other known failure modes or disadvantages? Er.... wait a minute. Tidy strips out the errors? So we do not serve sections/pieces of the information we intend to anybody? And this is desirable how? So we add to an existing known Wikimedia wide problem by adding local problems which can also be stripped out by Tidy. Is what Tidy delivers guaranteed to be W3C compliant or does it just address some problems to ease some of the pain for intrepid internet users attempting to access wikimedia based sites. Do the Wikimedia developers currently plan to get W3C html standard compliant in their project road map? If so does this give us (Wikiversity) a schedule deadline for having W3C compliant html? Is it impossible to achieve the desired aesthetics (rounded corners) with W3C standards compliant html? Mirwin 08:05, 8 February 2007 (UTC)

Which of those about a dozen questions do you want to have answered first? sebmol ^? 23:19, 8 February 2007 (UTC)

Any or all, if you have applicable information you feel like providing to the discussion. I am just trying to figure out how this conversation got to such an impasse with two Wikiversity administrators telling us that internet standards necessary to delivering quality content to all, not some, of the people on the planet are "trivial" and a matter of personal preference. If you are uninformed like me or feeling disinclined to share applicable data, do not sweat it. Other people can run googles on "Tidy"[1] if they feel like it or go talk to the technical developers if we/they decide we lack some useful expertise. If we decide we want to serve the entire planet well while continuing with local experimentation perhaps we can set up a sand box for you and Guillom to test out innovative experimental non working html. In fact I think someone posted a positive progress report towards that very thing for the Computer Science school at the Colloquium. In the meantime I will share my preliminary conclusion with you that David has presented an effective common sense solution to an urgent problem. Wikiversity is broken for many potential users. His solution: Leave the default so everybodys' W3C compliant browser works and coach people on how to modify their browser for rounded corners if they prefer that aesthetic. If this breaks their browser then at least they will be prepared to change back to a working configuration. A much better user experience than spending days trying to figure out why your system configuration does not work with Wikiversity only to find it was designed that way by local talent. Others wandering by will be capable of seeing the merit of David's proposal without fifteen or twenty years managing computer related engineering activities. It is not exactly rocket science. For his efforts David has been subjected to ridicule and belittlement apparently so a few of our administrators can have it their way rather than configuring our systems to server our users well. Mirwin 05:48, 9 February 2007 (UTC)

From the Tidy FAQ: "People continue to report examples where Tidy does not catch some ill-formed HTML or, worse, generates ill-formed HTML. These cases have been significantly reduced. That said, be sure to test Tidy with some representative files from your environment." They also suggest regression tests if the Tidy code is changed or html source is changed. Pretty typical or standard practice. Mirwin 07:51, 9 February 2007 (UTC)

Previous discussion on mainpage discussion page [2] Mirwin 00:12, 10 February 2007 (UTC)

I think their might be two debate going on here, whether to have rounded corners or not and whether to have code that renders differently from browser to browser. Is there anyway to get the rounded corners to show up on all browsers and operating systems? If there was, then we could concentrait on "fixing" the rounding code.--Rayc 20:16, 8 February 2007 (UTC)

Hi if people want rounded corners they could use this elegant solution for rounded corners in general and this for rounded tabs specifically. Both are cross browser compatible and W3C compliant. However I'd like to suggest we return to the default Monobook.css and put any changes in a separate style sheet such as Monobookrounded.css. Then people have a free choice of style to go with their free content. I'd like to add that I concur with what Mirwin has said above and in particular that David has been treated unfairly. Mystictim 15:41, 9 February 2007 (UTC)

Clarification for further discussion

After a discussion yesterday on IRC, I was asked to weigh in.

First, let me state that I personally like rounded corners, but overall, I'm not sure that this matters. Square corners would be fine with me, if that is what the consensus is. (I don't think this is the case, btw). But I'm not here to kindle a square vs. round debate.

I think we should focus on whether the code in question is W3C compliant, and what level of compliance we are looking to achieve, and how many browsers (and % of overall users) view the site the way we hope.

Our aim, whether it be corners or any other part of the site, should be to make the site functional for the highest % of users. This should mean that while the site should aim to look beautiful, as long as it degrades gracefully and functionality is preserved, that would be acceptable.

If the community wants round corners, and we can do this in a W3C compliant way, and this degrades gracefully on non-compliant browsers, then we should be doing it this way. Comments? Historybuff 01:55, 12 February 2007 (UTC)

I agree with you and Mystictim. Mirwin 04:23, 12 February 2007 (UTC)

I think that the concensus is to use W3C compliant code. Whether that be rounded or non-rounded is still no concensus. If the current code is not W3C compliant, it needs to be switched to Mystictim's code.--Rayc 18:59, 12 February 2007 (UTC)

More code solutions for rounded corners. Mystictim 20:16, 12 February 2007 (UTC)

Agreed. I never claimed that there was consensus that square corners are better, but there clearly is not consensus for this particular implementation of rounded corners. If there's a W3C-compliant alternative that would generate high-quality results in all compatible browsers, I don't object to its implementation. —David Levy 20:24, 12 February 2007 (UTC)

Comments

I've been holding back from this discussion, but I hope that we can come to a mutually satisfying resolution on the matter soon. Here are my comments for the moment on the issue of rounded corners, and this discussion in particular.

I've run the rounded corners by my girlfriend (a designer), where I showed her Wikiversity's main page in Safari and in Firefox and simply asked her if she saw anything different - she noticed the rounded corners immediately, and (unprompted) began criticising their jagged edges and lack of space around the top of the text. I've uploaded this, this and this to show how the rounded corners look on my computer (a Mac) with Firefox. I have to say, I also think they look cramped, jagged, and broken - I can't say if this would turn me away from Wikiversity as a new user, but it does look badly done.

However, David's basic point here (apart from a dislike of the rounded corners themselves) has been that this change was applied as the default option. I don't think we should be imposing by default an option that clearly annoys some people, and that isn't completely straightforward to resolve. Clearly, a small but steady number of people have voiced disapproval (even as a steady number have voiced approval) - this does not appear to me to be anything like "consensus" (our fundamental organising principle). Obviously, we encourage being bold - but if the edit is not undoable by a normal user, there needs to be either consensus behind that action or an attitude of "no big deal" to undo it. Furthermore, it should not be up to people who complain to create a non-standard page and then add lines of code to it. Many people might be annoyed by the corners, but will not complain - and of the people who take to time to find out how it can be resolved, many will just not do this, and they need to be logged in to do it in the first place. Wikis are intimidating enough to many people - and this doesn't seem like a very user/people-friendly process.

Following on from this, and although I don't have a great understanding of W3C compliant code/HTML, I would say that I think we should be particularly sensitive to usability-related issues on Wikiversity. Seeing as this is an educational website, it seems like the sensible thing to do. Rounded corners may not be the greatest impediment to someone navigating the site, but it doesn't seem to create a good precedent to create a site design that does not render consistently for the end-user.

For me, everyone here seems to agree that 1) rounded corners are a good idea in principle, and 2) this issue should not necessarily have given rise to such a big (and hot) debate. Therefore, I echo Mystictim's comment that we should revert back to the original default (ie square corners), and then work on a button layout that is, by all means, distinctive and attractive, but most importantly, acceptable to all (or as close as possible). I'd also like to reaffirm that we're building an educational community here, and that the way to do this, IMO, is through mutual respect. I don't think we'll get very far if we ignore this guiding principle - or if we cannot take, as well as give, criticism - or if we immediately treat a complaint as a threat. I think we've much to learn as we develop this project - and I feel we've something to learn here. Cormaggio ^beep 02:02, 13 February 2007 (UTC)

I prefer the rounded corners. --HappyCamper 03:45, 14 February 2007 (UTC)

:-)(Btw, did you view the images I provided?) But we will never get to the bottom of this discussion if it's as simple as one person's aesthetics against another's. The point, for me, is that a significant enough population seem to dislike the current rounded corners quite intensely. Perhaps if we put it to the vote, the rounded corners would have it - but that doesn't seem to me to give a statisfying, consensual result, in the light of what's been said against them. Cormaggio ^beep 13:03, 14 February 2007 (UTC)

Arguments such as it should "render consistently for the end-user" and it should be "W3C compliant" strike me as being just as subjective as "I do not like the jaggies". In my view, the small jaggies in the rounded corners are not bothersome enough to outweigh the positives of rounded corners, I don't understand why anyone cares if it renders consistently for all browsers and I don't know why anyone cares if the CSS code for it matches the W3C guidelines. I'm still waiting for someone to explain why all browsers must produce the same results or why it matters if the CSS code for the round corners is not in a W3C list of standard code. "dislike the current rounded corners quite intensely" <-- this seems get to the heart of the matter. There has been a long stream of unexplained exaggeration of the jaggie problem: "terrible", "horrible", "major problem", "extremely bad". I have tried to understand what might motivate the use of such terms, but after much careful examination of the matter I still do not understand how these terms apply to such a minor matter of taste. Exaggeration of of the facts does not create conditions for mutual respect. When much noise is made but no rational argument is offered I listen, discuss and then move on. Wikiversity should not degenerate into a playground where the loudest shouter wins the argument just because the playground monitor does not want to hear more shouting. I would object to putting the matter to a vote before a rational case has been made for taking action. I think the way to show that you "can take" criticism from an exaggerating complainer is to politely ask for a rational argument from the complainer. Until a rational argument is provided, I prefer to keep my attention centered on more important matters. --JWSchmidt 15:43, 14 February 2007 (UTC)

1. A consistent (or reasonably consistent) end-user interface is desirable because it allows us to create a good, distinctive appearance for all users (which persists even when they switch browsers). The converse (an interface deliberately designed to vary for no valid reason) is unprofessional and disconcerting.

2. Yes, the opinion that the rounded corners look terrible is entirely subjective. So what? So is your opinion that they look good (keeping in mind that they vary substantially from one computer to the next). You've unilaterally determined that people with whom you disagree are exaggerating, and you've cited this as justification for dismissing their viewpoints.

This is a purely aesthetic element, so the subjective belief that it looks bad is a rational argument. So is the subjective belief that it looks good, but there clearly is no consensus for that. As this is a nonstandard software hack (and no one has complained about the standard configuration's appearance), the onus is on your side to establish consensus for the rounded corners. If it seems to you like such a "minor matter of taste," you shouldn't mind deferring to those who obviously have far stronger opinions (perhaps because the effect looks worse on their displays than it does on yours). —David Levy 18:12, 14 February 2007 (UTC)

"interface deliberately designed to vary" <-- In situations where there are standards, people will experiment and do things in a non-standard way. Standards often change to follow practice. "no valid reason" <-- I have no idea if allowing some browsers to use this code was done for a "valid reason" or even what might constitute a valid reason in this situation. "unprofessional and disconcerting" <-- I switch between square corners and round corners every day. I do not understand how doing so would be disconcerting to anyone. I suppose many people think that being professional means placing the label "unprofessional" on every change that goes against conventional practice. I was trained in a profession where if you object to something you are expected to provide a rational reason for your objection. "subjective belief is a rational argument" <-- I would say that a rational argument can include balanced concern for subjective biases. "unilaterally determined that people with whom you disagree are exaggerating" <-- I have repeatedly expressed my concern about some of the the language that has been used to characterize the appearance of the rounded corners. I have suggested that many of the terms that have been used are not appropriate and that they are exaggerations. I have waited for others to explain why these questionable terms are appropriate. Since no justification has been provided for the use of terms like "terrible", "horrible", "major problem", "extremely bad" I must continue to express my feeling that these terms do overstate the jaggie problem and their use is an attempt to mis-characterize what is actually a very minor problem. I do not dismiss the fact that jaggies are annoying, but in trying to reach a decision that balances the pros and cons of the rounded corners I do try to dismiss irrelevant and misleading information. I you honestly feel that "terrible", "horrible", "major problem", "extremely bad" correctly characterize the rounded corners then we will just have to agree to disagree. "the onus is on your side to establish consensus" <-- We are experimenting with many new ways of doing things at Wikiversity. It is common knowledge that some people freak out when they see things being done in a new way. In many cases, if you give something that is new a trial period, you come to like it. Given these realities, it does not make sense to abandon experiments as soon as someone voices a doubt or a concern. Most people have not complained about the round corners and that silence is the best measure of consensus. "you shouldn't mind deferring to those who obviously have far stronger opinions" <-- Persuasive reasoning is much more likely to influence my actions than strong opinions. I've spent a huge amount of time examining the rounded corners in different operating systems, using different browsers, on various displays and at many resolutions. In every case the jaggie problem is a minor visual distraction. Yes, there is less area under the rounded buttons, but it is a very small decrease. --JWSchmidt 16:58, 15 February 2007 (UTC)

In situations where there are standards, people will experiment and do things in a non-standard way. Standards often change to follow practice.>

You're using "non-standard" to mean "out of the ordinary." I'm referring to international HTML coding rules (formalized by an official organization). They change over time, but this is an ongoing, small-scale experiment that has generated highly unsatisfactory results. (The code itself—not merely its use—is literally experimental.)

I have no idea if allowing some browsers to use this code was done for a "valid reason" or even what might constitute a valid reason in this situation.>

1. I meant that we have no valid reason to want to generate a substantially different interface from one graphical browser to the next.

2. No one sat down and decided to "[allow] some browsers to use this code." It works only in Mozilla browsers because it's a proprietary invention that Mozilla is experimenting with and no one else has adopted.

I switch between square corners and round corners every day. I do not understand how doing so would be disconcerting to anyone.>

You're familiar with the underlying cause. Most people aren't. When a website's interface continually changes back and forth for no apparent reason, people are distracted and believe that they've done something wrong or that someone's fooling around. Some seemingly innocuous visual changes can be caused by viruses, so someone unfamiliar with this Mozilla coding technique might even suspect that his/her computer has become infected with one.

I suppose many people think that being professional means placing the label "unprofessional" on every change that goes against conventional practice.>

No. It's unprofessional to use poorly implemented, experimental HTML code. And yes, this code is very poorly implemented. The fact that it produces dramatically variable results depending on one's operating system (despite the fact that the browser itself is the same) is indicative of that. This is something that could improve over time, but the Mozilla developers don't appear to even be working on it currently; it was thrown in for fun (and possible future development), not as a serious addition to the HTML library.

I was trained in a profession where if you object to something you are expected to provide a rational reason for your objection.>

Rational reasons have been provided, and you've unilaterally dismissed them.

I have repeatedly expressed my concern about some of the language that has been used to characterize the appearance of the rounded corners. I have suggested that many of the terms that have been used are not appropriate and that they are exaggerations.>

I, conversely, believe that these terms are not exaggerations. Why does your opinion outweigh mine?

I have waited for others to explain why these questionable terms are appropriate. Since no justification has been provided for the use of terms like "terrible", "horrible", "major problem", "extremely bad" I must continue to express my feeling that these terms do overstate the jaggie problem and their use is an attempt to mis-characterize what is actually a very minor problem.>

You've paradoxically deemed opinions wrong and demanded that people prove otherwise. This is impossible.

I do not dismiss the fact that jaggies are annoying, but in trying to reach a decision that balances the pros and cons of the rounded corners I do try to dismiss irrelevant and misleading information.>

You've continually referenced the "pros," but I've yet to see you cite a single one (other than "I like the rounded corners"). Meanwhile, you've decided to "dismiss" the feedback of people with whom you disagree.

I you honestly feel that "terrible", "horrible", "major problem", "extremely bad" correctly characterize the rounded corners then we will just have to agree to disagree.>

Quite right. As there clearly is no agreement on this matter, please remove the non-consensus code.

We are experimenting with many new ways of doing things at Wikiversity. It is common knowledge that some people freak out when they see things being done in a new way. In many cases, if you give something that is new a trial period, you come to like it. Given these realities, it does not make sense to abandon experiments as soon as someone voices a doubt or a concern.>

1. You're acting as though this is a new teaching method (or something along those lines). Again, it's a purely aesthetic matter. People either like the rounded corners or they don't, and there's no consensus in favor of them.

2. This code is not "new." It's been around far longer than Wikiversity has, it's never improved or grown beyond its experimental state, and it's been widely criticised from day one.

Most people have not complained about the round corners and that silence is the best measure of consensus.>

Do you see dozens of people lining up to praise the rounded corners? There has not been substantially more support than opposition. That isn't "consensus."

As I explained, only ~12.5% of users even see the rounded corners. The remaining users (~87.5%) are unaware of this element's existence, so their silence cannot be construed as tacit approval.

Persuasive reasoning is much more likely to influence my actions than strong opinions.>

You should be persuaded by the fact that there's no consensus. Instead, you're demanding that people prove your opinions wrong (an impossibility).

I've spent a huge amount of time examining the rounded corners in different operating systems, using different browsers, on various displays and at many resolutions. In every case the jaggie problem is a minor visual distraction.>

To you, it's minor. To me, it's major. Neither your opinion nor mine is sacrosanct. There is no consensus.

Incidentally, I'll ask you to kindly explain how even a "minor visual distraction" is conducive to learning.

Yes, there is less area under the rounded buttons, but it is a very small decrease.>

In Cormac's screen captures, this difference is quite large. That's just one of the inexplicably unpredictable elements that render this code unacceptable for mainstream use. Who knows what other users are seeing? Who knows how many people quickly passed through Wikiversity because they believed that its coding was broken and didn't take it seriously? —David Levy 14:27, 16 February 2007 (UTC)

"Rational reasons have been provided, and you've unilaterally dismissed them." <-- My problem is that I am struggling to deal with a characterization of the problem with the rounded corners that does not mesh with reality as I experience it. In my experience, any "terrible", "horrible", "major" or "extremely bad" problem with a wiki user interface is quickly recognized and fixed, almost always by the person who introduces the problem. On rare occasions a serious problem might initially escape notice, but upon being noticed and discussed a serious problem will then be quickly fixed. "You've paradoxically deemed opinions wrong" <-- Terms such as "wrong" and "bad" are not terms I like to apply to subjective opinions. I'd rather say that I find it inexplicable that anyone would apply terms like "terrible", "horrible" and "extremely bad" to a minor case of jaggies. "You've continually referenced the pros, but I've yet to see you cite a single one" <-- I have to preface my comments with a disclaimer: my preference for the rounded corners is mainly a simple matter of taste. I can produce rationalizations, but I do not trust them very much. I liked the rounded corners as soon as I saw them, even before I thought about the rationalizations. However, the rationalizations that do seem to influence me are these:

• I think the curved tabs at the top of the page complement the Wikiversity logo.
• Having part of the Wikiversity interface look different (compared to the English Language Wikiversity Wikipedia) fits nicely with Wikiversity being a place where experiments are being done as part of an exploration of how to use wiki technology in new ways. Wikiversity might be a little rough around the edges but we are trying to get people to think outside the box.
  

"you've decided to dismiss the feedback of people with whom you disagree" <-- I'd rather characterize the situation as my failing to be impressed or moved by comments that I do not understand. "please remove the non-consensus code" <-- I will not. I'm still not moved. "there's no consensus in favor" <-- Most people have said nothing. I take that silence as indicating that there is no need to remove the code for the rounded corners. "demanding that people prove your opinions wrong" <-- I'm not sure that is a constructive way of characterizing what has been going on here. I try to avoid casual use of the words "prove" and "proof". I try to keep clearly in mind the distinctions between subjective opinion and objectively verifiable evidence. "explain how even a minor visual distraction is conducive to learning" <-- A good lesson is learned when people come to recognize the things in this world that are important and ignore things that are minor annoyances. On the positive side, I take seriously the idea that the rounded corners can be a visual metaphor for a spirit of experimentation, a spirit that is useful to learners. "less area under the rounded buttons...this difference is quite large" <-- I suppose we could count the number of pixels inside a rectangular tab and compare that to a rounded tab. My eye tells me that it is a small difference. "Who knows how many people quickly passed through Wikiversity because they believed that its coding was broken and didn't take it seriously?" <-- I do not know how many, but my guess is "zero". --JWSchmidt 04:26, 17 February 2007 (UTC)

My problem is that I am struggling to deal with a characterization of the problem with the rounded corners that does not mesh with reality as I experience it.>

The problem is that you refuse to accept that this is a matter of opinion.

Terms such as "wrong" and "bad" are not terms I like to apply to subjective opinions. I'd rather say that I find it inexplicable that anyone would apply terms like "terrible", "horrible" and "extremely bad" to a minor case of jaggies.>

To you, the jaggies seem "minor." I find it inexplicable that you would be unable to realize that rational humans are capable of legitimately disagreeing with your assessment.

I have to preface my comments with a disclaimer: my preference for the rounded corners is mainly a simple matter of taste.>

Of course it is! Others have cited our personal preference for the straight corners. We also have cited technical and sociological reasons why the use of unpredictable, nonstandard, experimental HTML code is a bad idea.

I think the curved tabs at the top of the page complement the Wikiversity logo.>

I disagree. I think that the straight-cornered tabs complement the Wikiversity logo.

Having part of the Wikiversity interface look different (compared to the English Language Wikiversity) fits nicely with Wikiversity being a place where experiments are being done as part of an exploration of how to use wiki technology in new ways.>

1. I assume that "English Language Wikiversity" is a typo for "English Language Wikipedia."

2. No one has objected to the idea of creating a distinctive appearance for Wikiversity. Unfortunately, this code deliberately prevents us from having a consistent user interface. How can we create a coherent overall design when different people see dramatically different things?

Wikiversity might be a little rough around the edges but we are trying to get people to think outside the box.>

Does "think[ing] outside the box" include include forcing non-consensus software hacks on the community?

I'd rather characterize the situation as my failing to be impressed or moved by comments that I do not understand.>

You don't understand how a rational human could possibly disagree with you on a completely subjective matter?

I will not [remove the code]. I'm still not moved.>

I'm not asking you to alter your personal opinion. I'm asking you to honor the fact that there is no consensus for this hack.

Most people have said nothing. I take that silence as indicating that there is no need to remove the code for the rounded corners.>

1. Again, "most people" cannot see the rounded corners, so how do you expect them the comment on them?

2. Again, among the people who have commented, there is no consensus. Your requirement that we hear from "most people" (a minimum of 3,337 users, excluding those who haven't registered accounts) has absolutely no basis in policy.

I'm not sure that ["demanding that people prove your opinions wrong"] is a constructive way of characterizing what has been going on here. I try to avoid casual use of the words "prove" and "proof". I try to keep clearly in mind the distinctions between subjective opinion and objectively verifiable evidence.>

Then why are you insisting that users somehow substantiate their subjective opinion that the rounded corners look extremely bad and refute your subjective opinion that this is an exaggeration?

A good lesson is learned when people come to recognize the things in this world that are important and ignore things that are minor annoyances.>

Wow. You're actually arguing that the deliberate insertion of minor (in your assessment) annoyances is beneficial to the community. Wow.

On the positive side, I take seriously the idea that the rounded corners can be a visual metaphor for a spirit of experimentation, a spirit that is useful to learners.>

Do you take seriously the spirit of consensus-based decision-making and your duty as a sysop to uphold this principle?

I suppose we could count the number of pixels inside a rectangular tab and compare that to a rounded tab. My eye tells me that it is a small difference.>

Did you look at Cormac's screen captures? Do you understand that the appearance varies from computer to computer? —David Levy 06:38, 17 February 2007 (UTC)

You wrote "Do you take seriously the spirit of consensus-based decision-making and your duty as a sysop to uphold this principle?" Consensus-based decision-making is but one method of making decisions, it's not necessarily the way to make decisions at Wikiversity. You again extrapolate from your experiences of hwo things work on Wikipedia to how things should work on Wikiversity. sebmol ^? 11:10, 17 February 2007 (UTC)

Oh dear. Sebmol, surely consensus is our fundamental organising principle. Of course, people can be bold and decide to do whatever they want. But consensus is the only way of ensuring that people feel able to express themselves, and that their opinions can be genuinely taken into consideration - newcomers and oldtimers alike.

In fact, this is pretty much what I was going to write about before Sebmol posted - genuinely taking someone's opinion into consideration. As JWSchmidt says, we are conducting an experiment here with the rounded corners. However, this experiment has been successful for many, but not for some. It seems to me that we need to understand the reasons for the experiment's lack of success (for some), and improve upon the experiment. For me, the corners looking jagged and squashed is bad, but worse is the fact that the text is given no space in the boxes. Mike suggests below that there's very little that could be done with so many pixels, but could at least this much be addressed? And what about the other code suggestions made above by Mystictim? Could someone css-savvy enough try to find a solution here, instead of failing to recognise that there is a problem? Cormaggio ^beep 12:24, 17 February 2007 (UTC)

Firstly, thank you, Cormac.

Michael's opinion presumably is derived from the fact that he happens to have a configuration (some combination of factors) that causes the rounded corners to look smooth on his display.

The visual quality problem stems from the fact that this code has been very poorly implemented at the browser level (by Mozilla developers who threw it in for fun and have left it sitting on the back burner); there is nothing that we can do to make this code work better for anyone.

It might be possible to replace the Mozilla code with different code that generates high-quality rounded corners for everyone with W3C-compliant browsers. Someone more knowledgeable with the alternative methods could comment on whether it's feasible to integrate them into MediaWiki.

In the meantime, there clearly is no consensus for the method in use, and I'm more than mildly surprised to see a sysop and a bureaucrat arguing that this is irrelevant at Wikiversity. —David Levy 17:43, 17 February 2007 (UTC)

Consensus-based decision-making is but one method of making decisions, it's not necessarily the way to make decisions at Wikiversity.>

So...what is the way to make decisions here? Whoever has the most buttons wins?

You again extrapolate from your experiences of hwo things work on Wikipedia to how things should work on Wikiversity.>

Actually, Sebmol, I was relying on my experience as a Wikimedian and Cormac's comments from the beginning of this section (which you might want to read). —David Levy 17:43, 17 February 2007 (UTC)

"you refuse to accept that this is a matter of opinion" <-- I find no way to avoid characterizing this as a matter of opinion vs exaggeration. "I find it inexplicable that you would be unable to realize that rational humans are capable of legitimately disagreeing with your assessment" <-- exaggeration is not a method of "legitimately disagreeing", and I've explained the objective criteria by which I characterize this endless discussion as having being built upon exaggeration. "unpredictable, nonstandard, experimental HTML code is a bad idea" <-- the code for rounded corners produces results that are well understood. In my opinion, what you call "bad" is a minor annoyance. "this code deliberately prevents us from having a consistent user interface. How can we create a coherent overall design when different people see dramatically different things?" <-- I'm not concerned about the fact that only some browsers show the rounded corners. "forcing non-consensus software hacks on the community" <-- I'm not sure it is constructive to try to characterize this little experiment with rounded corners in such harsh terms. "You don't understand how a rational human could possibly disagree with you on a completely subjective matter" <-- I do not recognize my position as being correctly characterize by this statement."the fact that there is no consensus for this hack" <-- One person using exaggerated claims to endlessly push one POV does not enter into my estimation of consensus. "most people cannot see the rounded corners, so how do you expect them the comment on them?" <-- I never said that I have such an expectation. "Your requirement that we hear from most people" <-- I'm not familiar with the existence of any such "requirement". "why are you insisting that users somehow substantiate their subjective opinion that the rounded corners look extremely bad" <-- I've explained previously that in my experience, when there is a problem with a wiki user interface that is "extremely bad" then nobody has to be a lone voice calling in the wilderness in order to fix the problem. I take that as an objective test for the presence of a problem that is "extremely bad". "Do you take seriously the spirit of consensus-based decision-making and your duty as a sysop to uphold this principle?" <-- Yes. "Did you look at Cormac's screen captures?" <-- Yes "Do you understand that the appearance varies from computer to computer?" <-- Yes, I previously explained my personal experiences with variation in the jaggie problem under various conditions. --JWSchmidt 04:23, 18 February 2007 (UTC)

I find no way to avoid characterizing this as a matter of opinion vs exaggeration.>

Whether something looks "good" or "bad" is a matter of opinion. You are not the arbiter of what opinions constitute "exaggeration."

exaggeration is not a method of "legitimately disagreeing", and I've explained the objective criteria by which I characterize this endless discussion as having being built upon exaggeration.>

You've basically said, "it doesn't look so bad to me," and cited this as evidence that we're wrong.

the code for rounded corners produces results that are well understood.>

Yes, we understand that the results vary wildly from one computer to the next, look extremely poor to many people, and fail to manifest at all for most.

In my opinion, what you call "bad" is a minor annoyance.>

Yes, in your opinion. There is no consensus for your opinion.

I'm not concerned about the fact that only some browsers show the rounded corners.>

Other people are. Why does your opinion outweigh ours?

I'm not sure it is constructive to try to characterize this little experiment with rounded corners in such harsh terms [as "forcing non-consensus software hacks on the community"].>

You refuse to remove a software hack that lacks consensus, so I don't know of a better way to describe the situation.

One person using exaggerated claims to endlessly push one POV does not enter into my estimation of consensus.>

1. One person? What are you talking about?

2. For the sake of argument, let's assume that the criticisms are exaggerated. What difference does it make? If you ignore the complaints' severity and simply count them as expressions of preference for the straight corners, there is no consensus for the rounded corners.

In actuality, of course, these comments about the rounded corners are complaints (and I've yet to see anyone from any wiki complain about the straight corners).

I never said that I have such an expectation [to see most people comment on the rounded corners].>

I'm not familiar with the existence of any such "requirement" [that we hear from most people].>

You stated that "most people have not complained about the round corners and that silence is the best measure of consensus." How else, other than a demand that "most people" complain before you'll act, can this be interpreted?

I've explained previously that in my experience, when there is a problem with a wiki user interface that is "extremely bad" then nobody has to be a lone voice calling in the wilderness in order to fix the problem.>

I'm not a lone voice. I'm (evidently) the one lowly peon with the patience to persist in debating this matter with sysops who believe that their positions give them the authority to overrule the opinions of others. —David Levy 06:57, 18 February 2007 (UTC)

"You are not the arbiter of what opinions constitute exaggeration" <-- I've described the objective metric I use for judging the severity of the rounded corners problem. "You've basically said, 'it doesn't look so bad to me,' and cited this as evidence that we're wrong." <-- I apply objective criteria to judge if my perceptions are correct about the exaggerations. "the results vary wildly from one computer to the next" <-- I do not see that. How does the term "wildly" apply?. "Why does your opinion outweigh ours?" <-- I've described the objective metric by which I judge exaggerated claims about the severity of the rounded corners problem. "How else, other than a demand that most people complain before you'll act, can this be interpreted?" <-- If there were a serious problem with the rounded corners it would be evident to everyone and it would have been fixed long ago. If it were a serious problem, it would have been fixed long before many people would have had to complain. "I'm (evidently) the one lowly peon with the patience to persist in debating this matter with sysops who believe that their positions give them the authority to overrule the opinions of others" <-- I have no problem ignoring exaggerated opinions. --JWSchmidt 08:50, 18 February 2007 (UTC)

I've described the objective metric I use for judging the severity of the rounded corners problem.>

I apply objective criteria to judge if my perceptions are correct about the exaggerations.>

I've described the objective metric by which I judge exaggerated claims about the severity of the rounded corners problem.>

No. You've described the criteria by which you formulate your subjective opinions. I could supply a similar list of criteria by which I evaluate a painting's beauty or a cake's flavor, but that wouldn't make me "correct" and other people "incorrect." If I like the cake, but half of the party guests declare that it tastes "terribly over-sweet," I may believe that they're exaggerating, but that doesn't mean that anyone is right or wrong. The only "objective metric" is consensus. There is no consensus for the rounded corners. Again (as you ignored this part of my previous reply), if you disregard the issue of severity and simply divide the respondents into a "prefers the rounded corners camp" and a "prefers the straight corners camp," there is no consensus. I don't understand how you can possibly justify leaving the non-consensus hack in place (instead of restoring the default appearance that no one has complained about).

Incidentally, the third quote addressed my question, "Why does your opinion outweigh ours?", which I wrote in response to your statement that you're "not concerned about the fact that only some browsers show the rounded corners." You replied out-of-context (as though my question pertained to the "exaggerated" claims of ugliness). Again, why does your lack of concern regarding cross-browser compatibility outweigh others' concern? What gives you the authority to overrule contrary viewpoints?

I do not see that. How does the term "wildly" apply?>

Cormac's screen captures (which show something significantly different from the results that I obtain via my computer) and comments from other people about obtaining still other results are evidence of a seemingly wild (id est untrammeled) degree of variance from one computer to the next. How can you dispute this (even if you believe that it isn't a big deal)?

If there were a serious problem with the rounded corners it would be evident to everyone and it would have been fixed long ago.>

Oh, so now "everyone" has to agree before these concerns are taken seriously? You claimed to respect the spirit of consensus-based decision-making, and now you're demanding unanimity. Again, wow.

If it were a serious problem, it would have been fixed long before many people would have had to complain.>

You're confusing the concept of a "terrible problem" with that of a "problem in which something looks terrible." When a cake tastes "terrible" to a significant number of party guests, that isn't a "terrible" occurrence (in the grand scheme of things). It's only cake, after all. Nonetheless, it's a problem that's addressed by not serving that type of cake again.

No one is claiming that this issue will lead to Wikiversity's downfall. We're noting that there is no consensus for the "rounded corners" hack, so the only proper course of action is to remove it.

I have no problem ignoring exaggerated opinions.>

You have no problem ignoring the opinions of people who disagree with you and forcing them to accept a non-consensus hack that you happen to like. That's rather disconcerting. —David Levy 18:14, 18 February 2007 (UTC)

"The only objective metric is consensus" <-- I will continue to ignore information that does not help me to measure consensus. "I don't understand how you can possibly justify leaving the non-consensus hack in place" <-- In my judgment, the fact that most people do not complain means that there is no need to change the rounded corners. "why does your lack of concern regarding cross-browser compatibility outweigh others' concern?" <-- I'm still waiting to hear a reason that will make me care that only some browsers show the rounded corners. "wild degree of variance from one computer to the next....How can you dispute this" <-- All the results I have seen look about the same. I see nothing I would call "wild". "so now everyone has to agree before these concerns are taken seriously?"<-- I never said that. "You're confusing the concept of a "terrible problem" with that of a "problem in which something looks terrible." <-- I'm not confused. "there is no consensus for the rounded corners" <-- We are performing an experiment. I take the fact that most people do not complain about the rounded corners to be the most significant information available. "You have no problem ignoring the opinions of people who disagree with you" <-- I have no problem ignoring exaggerated opinions. --JWSchmidt 19:35, 18 February 2007 (UTC)

I will continue to ignore information that does not help me to measure consensus.>

Consensus is not measured by ignoring the subjective opinions of people with whom you disagree. But again, even if you ignore the severity of the complaints (and interpret them as comments along the lines of "I prefer the straight corners"), there is no consensus.

In my judgment, the fact that most people do not complain means that there is no need to change the rounded corners.>

Okay, so now you're back to saying that "most people" (3,363 users, excluding those who haven't registered accounts) would have to complain about the rounded corners before you'd see justification to remove them. And again, you're ignoring the fact that most people cannot see the rounded corners and the fact that of those who have commented, there is no consensus.

I'm still waiting to hear a reason that will make me care that only some browsers show the rounded corners.>

Your opinion is not sacrosanct. You don't have to care. You only have to honor the fact that other people care. "Convincing JWSchmidt that something is right" is not a criterion in the consensus process.

All the results I have seen look about the same.>

Really? My screen captures look about the same to you as Cormac's? I couldn't disagree more.

I never said that [everyone has to agree before these concerns will be taken seriously].>

You stated that "if there were a serious problem with the rounded corners it would be evident to everyone and it would have been fixed long ago" (emphasis mine).

We are performing an experiment.>

There is no consensus that this "experiment" should continue. Your extra buttons don't give you the right to force your personal preference on the community.

I take the fact that most people do not complain about the rounded corners to be the most significant information available.>

That isn't how "consensus" works. If we waited for "most people" to comment before making decisions, no decisions would ever be made. Consensus (or the lack thereof) is determined via the opinions of whichever people do comment. As such, there is no consensus for this hack.

I have no problem ignoring exaggerated opinions.>

And for some reason, you believe that you possess the authority to unilaterally determine whose opinions are "exaggerated" (which happens to be everyone with whom you disagree). —David Levy 20:23, 18 February 2007 (UTC)

---

Which browsers render the rounded corners smoothly and which render the rounded corners jagged?--mikeu 18:20, 16 February 2007 (UTC)

Only Mozilla browsers (with Firefox being the most popular) render the rounded corners at all. Their specific quality appears to be dependant on various factors (operating system, resolution, display type, et cetera). —David Levy 18:52, 16 February 2007 (UTC)

Not just Mozilla browsers, I'd imagine all gecko-based ones would also. (and all Mozilla browsers use gecko). Anyway, just to put in my 2 cents - at least on my configuration - I'm at a loss as to how corners could be more round in so few pixels. Michael Billington (talk • contribs) 04:59, 17 February 2007 (UTC)

1. I was referring to the Mozilla Foundation (hence the link). The Gecko engine is developed via the Mozilla Foundation.

2. We've established that the rounded corners look good under some configurations. You evidently possess one. —David Levy 06:38, 17 February 2007 (UTC)

Going back to go forward

From this page and two other relevant discussions (here and here), I count the following opinions:

• 8-10 people (variance is in case anonymous editors also participated as logged-in users) who dislike rounded corners: David Levy, Robchurch, Mikeu (Mu301), Zioroboco, Mystictim, Mirwin, myself (+ girlfriend), and two anonymous editors, 85.147.58.212 and 83.78.187.95
• 7 people who like the rounded corners: Sebmol, Robert Elliott, Guillom, JWSchmidt, HappyCamper, HistoryBuff, and Mike Billington.

This is not (as is claimed) a "lone voice in the wilderness" who disagree with the up-to-current implementation of rounded corners - on the contrary. There is no discernible general approval for this change - some people like it, and some people dislike it. It seems easier for people who like the code to actively place this code in their own Monobooks than to expect someone who doesn't like it to figure out how to change it. For this reason alone (but also combined with the mini-stats above), I feel justified in reverting back to having square corners as the default option.

However, I do not want to undermine the efforts (and yes, experiments) to make Wikiversity more attractive for its users. I personally would very much like corners and boxes to be round - if they did not render so inconsistently and (in many people's opinions) badly. I'd like to place the emphasis on improving what we have - perhaps going to solutions proposed here by Mystictim (which, unfortunately, rely on images) - in order to create something which is attractive, reasonably consistent (ie, quoting Historybuff, "degrading gracefully" - not notably worse), and which complies with good practice for web accessibility. I would note here that, as far as I can make out (and I've spent a couple of hours reading accessibility-related websites), the rounded corners here don't break the W3C's major criteria for accessibility but that they do degrade badly enough for some people to notice and be put off by. One of the most obvious processes for assessing accessibility is to test it with people - and I would say that the up-to-current solution has clearly not passed the people test. So, let's stop arguing - and get on with actually improving this experiment - learning about what works and what doesn't work. Cormaggio ^beep 02:18, 19 February 2007 (UTC)

I've made a quick guide to adding rounded corners to one's own skin, see User:MichaelBillington/Rounded corners. Michael Billington (talk • contribs) 02:55, 19 February 2007 (UTC)

Thank you, Cormac! I hope that the proponents of the rounded corners can look beyond their personal opinions and recognize the lack of consensus. By the same token, while I actually like the straight corners, I'll gladly accept an alternative method of rounding the corners if one backed by consensus can be devised. It isn't the style that's the problem; it's the implementation. —David Levy 03:48, 19 February 2007 (UTC)

For the record I dislike experimental malfunctioning code on operational servers. I happen to like rounded corners and will probably eventualy implement should clean code become available. Mirwin 00:22, 20 February 2007 (UTC)

Yes, I'm sorry, I meant "dislike the rounded corners as have been implemented so far". I don't think anyone ever dismissed round corners in principle - and many (if not most) would like to see a solution that's better than what we have had so far. Cormaggio ^beep 01:22, 20 February 2007 (UTC)