# Difference between revisions of "ConTeXt wiki:Wishlist (old)"

Here's a wish list for the gardner :)

## Ideas

... a place to collect ideas to make the garden even better ...

### wiki

• Q: I have been playing a bit with the <context> rendering. What do you think of [1] instead of [2]? --pg
A: I'm not sure yet; it looks sharper, but the sharpening seems a bit uneven compared to the old version. So I don't know whether I prefer "sharper" or "more even". ... After looking at this some more, I think I slightly favor the old version, because it does look more even. On the other hand, the images are a lot smaller on the new version, which is a definite boon on a modem connection! --Brooks
• Q: adding <xmlcode> (code highlighter is already written) or, if not too complicated, something more general like <code type="xml">, <code type="mp">, ...
A: This is quite hard I'd guess. Addition: in the new wiki software which I will eventually install (I am currently testing a lot -- MW sofware doesn't behave as expected) will be able to handle those parameters (note for myself: instructions here)
• Q: adding <latex> counterpart to <context>; mostly in order to be able to compare LaTeX and ConTeXt results or to be able to describe the desired result in Category:From LaTeX
A: [update] I didn't remember what <context> does... Would we need a multiple pass solution? --pg
A2: Brooks seems to like LaTeX, perhaps he can tell more; I don't really remember how many time one needs to run it. I guess that we won't need BibTeX, but perhaps some proper indexing/cross referencing would be handy. Is running it two times enough? More than three times is probably already to much. --Mojca
A3:In most cases, LaTeX really only needs one run. It needs two runs to resolve any references that are referenced before they are created -- which, for Wiki examples (which don't have floats floating to later pages) is probably likely to be so infrequent as to be unnecessary. It occasionally needs more than two runs, but only in really rare cases, and we definitely won't see those on the Wiki. So, my answer would be to do two-pass (call "latex filename" twice) if it's easy to do, otherwise one-pass is probably sufficient. --Brooks 06:10, 9 Sep 2005 (UTC)

### archive

• Q: Is it possible to download attachments? A: I have 'disabled' (by removing the link) downloading attachments. Lurker offers part of the message as an attachment for download. This might include email addresses and I don't want to reveal them without changing. I fear that I would also change the integrity of the attachments with my perlscript. So: they are there but not linked. I'll write this down on my todo list.
• Q: I vaguely remember you were talking about using some other program for the html-ization (instead of / next to Lurker). Is that likely to happen any time soon? A: Not in the next two months. But I think around that time: yes. It will not use lurker as a backend

### articles

(Comment from the gardener: articles will be the first service which will change completely. I will try to use ruby on rails for that one. Thanks for the suggestions, I will implement them --pg )

• archiving articles (if allowed) or at least adding direct URLs to them
• sort by author/magazine number/title
• a form to submit proposals for new entries
• maybe: adding talks in conferences (EuroTeX, ...)
• unimportant: exporting bibtex entries about the magazine (in plain and Taco's format)

### live

• selector to choose the ConTeXt version

### source

• adding <a name="line xxx"> at the beginning of every line in the source code as to be able to reference to a certain line in hyperlinks (on the other hand: it may be a bad idea to be able to point to a certain line since the files are a subject to change)
• perhaps adding <a name="ccedilla"> to places where \definecharacter stands in a similar way as the alt name= are added in front of \def\somecommand (on the other hand: a bad idea since there may be more than two definitions of a single named character in one file)
• serching/highlighting more that one word
• if the user comes from google, parse the URL for &q=somequery and add ?search=somequery at the end of the new URL
Example: if someone searches for hack in google and follows the link to supp-vis.tex, the site http://source.contextgarden.net/tex/context/base/supp-vis.tex?search=hack would open instead

### general

<form method="get" action="http://www.google.com/search">
<table>
<tr>
<td>
<input name="q" size="31" maxlength="255" value="" type="text">
<input name="btnG" value="Search" type="submit">
<input name="sitesearch" value="contextgarden.net" type="hidden">
</td>
</tr>
</table>
</form>
The first two are implemented, see the help page. The third, well, needs some thinking and the link to the mailarchive will have to wait until I implement the new mailarchive. --pg

## Ideas: Done

### wiki

• Q: In the "recent changes" page, it's often hard to notice the "N" designations ("New page") in the midst of a long line of "M" ("Minor edit") changes. Would it be possible to highlight the "N" in red or something? --Brooks
I agree with it. It is possible to hide minor edits, but it would really be better if M would become M or if the color of N would change. --Mojca A: Done --pg

## Ideas: Rejected

### wiki

• Q: Are the <context> images rendered using gs? In that case, could -sDEVICE=pngalpha be used to get transparent images? A: [update2] Transparent images don't look right in IE, so I reverted the change. (Note for myself/1.5 wiki: see js at [3] for a fix. Thanks Mojca for this hint.)