User talk:Proteins/Accessibility archive

HTML lists

The HTML lists on your user page don't comply with the accessibility guidelines. In the list of ideas you're working on, each item is separated by a line break, so I hear: "list of 1 items, <list item>, list end, list of 1 items ...". The list of scripts is a little harder to fix: the colons that are used for indentation actually create tags that are used to make defenition lists, so they sometimes read strangely with JAWS. I think the most accessible way to fix this would be to use nested lists ("*" and "**"). This would be harder though because there are paragraphs in that section ... maybe put each script into a subheading? I don't know which way of formatting would look good while still being accessible. Perhaps the article structure script should check for lists where the items are separated by line breaks. Graham87 07:00, 22 October 2008 (UTC)

Sorry about that. I became aware of how lists and indented text are handled in Wikipedia only recently, when I was working on the articlestructure.js script. I fixed the lists on my user page as you suggested, and I'll try to improve the striplinks script to eliminate adjacent opening and closing tags of the same list type, such as "</ul><ul>". I'll let you know whether it works; there might be places where that elimination would be inappropriate. Proteins (talk) 12:01, 22 October 2008 (UTC)

Hi Graham, I think I've got the list-merger code working, for all three types of lists: unordered, ordered and discursive. The last one might be helpful in scanning through Talk pages, I imagine. I'd like to replace those pseudo-discursive lists used for indentation with some text such as "Indent level 1", "Indent level 2", etc. Unfortunately, there doesn't seem to be a way for the code to distinguish between those DL's and genuine discursive lists. If you can think of a way, let me know and I'll code it up.

I tried the revised script out on Lists of French people and on the FAC Talk page, where it seemed to work OK. If you have time, I would appreciate it if you would test the striplinks script out on a few pages and look for bugs. I also left a few unmerged lists on my user page. Thanks! Proteins (talk) 16:42, 22 October 2008 (UTC)

P.S. I changed the script to keep links to user names and user talk pages, and also to list all the redlinks that it removes, which seemed like it might be helpful to editors.

P.P.S. I studied the DOM structure of this very section, and it seems as though the script is not doing a perfect job of merging indented dialogues when the level of indentation goes beyond the first. For example, these last four paragraphs, which are doubly indented, do not merge into a single discursive list as they should; rather, they're elements within the outer discursive list. I'll work on fixing that. Proteins (talk) 16:51, 22 October 2008 (UTC)

The only difference between Wikipedia indents and proper definition lists is that Wikipedia indents don't contain a <dt> tag where as proper definition lists do. A good test of a script that deals with user and user talk pages is a request for adminship, and I've chosen Wikipedia:Requests for adminship/Jbmurray because it had a lot of people commenting. The edit section link isn't removed from the support section, and some of the user sig links (e.g. support #4) aren't linked. I'm thinking that the best use of the strip links script in talk pages would be to make indentation more clear, and there's not much point in removing links in talk page discussions, requests for adminship or articles for deletion discussions. Most links in talk page discussions are important, and there aren't too many of them compared to articles. The most annoying things for me on talk pages are templates like {{Userlinks}}, which I have to arrow past because there's no way of getting past it otherwise, and sometimes figuring out what the indentation level is. Graham87 08:17, 23 October 2008 (UTC)

Thanks, I've begun a separate script User:Proteins/unindent.js to take care of unindenting. It's beginning to work but I'm having trouble with places where people put linespaces between indented text. If you wish, you can try it out to see if you like the overall approach; or you can wait until it's working better. There are definitely bugs left. Proteins (talk) 15:17, 23 October 2008 (UTC)

The script works well at identifying the level of indentation for any indented text, and adds a prefix text to say what level it is. Unfortunately, the script can change the order of indented sections if the indented comments are not separated by a line space. More specifically, changes in order happen in unspaced multiply-indented sections when the indentation level reverses, for example, going from 1 to 2 to 3 then back up to 2. You can see examples at WT:FAC, but a simpler example is at this sandbox. The problem results from the way in which I outdent the text; I'm working on it. Proteins (talk) 22:03, 23 October 2008 (UTC)

Sounds good so far. It's just that the text "indent level 1" is too verbose for my liking - I'd prefer "indent 1" or "level 1" myself. I'd also prefer it if the indentation was only indicated if it changed - if someone writes a huge message with many paragraphs in a certain indent level, I'd prefer not to hear about it all the time. When people unindent, they usually indicate it in a discussion, so I don't think it's as important to indicate that. Graham87 01:04, 24 October 2008 (UTC)

I shortened the text to "Indent"+number, and I also fixed a minor problem, namely that if you ran the script twice on the same article, you would get multiple copies of the indent message and the color highlighting would be lost, since they had all been outdented to level one in the first run of the script. I'll work on the repetition problem next. But you need to watch out for the re-ordering of the sections! That's still a bug, and a bad one. You'll know when it happens because comments will be indented more deeply than the first level, the conversations will be unintelligible or out of logical order. I've been working on a new script User:Proteins/outdent.js that takes a new approach, but it still has bugs of its own. Proteins (talk) 17:57, 24 October 2008 (UTC)

Good news, I fixed the bugs in User:Proteins/outdent.js. You should definitely use only that script from now on, and not User:Proteins/unindent.js. I may still refine the script, but it seems to be working fine on my tests, including WT:FAC and Wikipedia:Requests for adminship/Jbmurray. Please try it out on your favorite pages and let me know if you catch any bugs!

One refinement I haven't included in this version is to eliminate the "Indent x" label in successive paragraphs indented at the same level. The coding would be more complicated and bug-prone, unfortunately. Let's get the script working as is, and then we can think about how to make that refinement on a bug-free script. Thank you, Proteins (talk) 23:00, 24 October 2008 (UTC)