255398 - create CSS Support Chart

Status: RESOLVED FIXED

(Developer Documentation Graveyard :: General, defect)

Description

[Richard York]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a2) Gecko/20040714
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a2) Gecko/20040714

I'd like to see Mozilla.org roll out with a chart of the CSS implemented in its
browser. Perhaps one may already exist, if so, please excuse me, I spent quite
some time looking for such a list but never found one.

Similar resources already exist for other browsers..
http://developer.apple.com/internet/safari/safari_css.html
http://www.opera.com/docs/specs/
http://msdn.microsoft.com/library/default.asp?url=/workshop/author/css/reference/attributes.asp

I'd be thrilled to compile such a list, but wouldn't know where to start! Given
the plethora of proprietary CSS properties, pseudo widgets, etc. Some direction
to the relevant source code libraries would help if such a list is deemed useful.

Reproducible: Always
Steps to Reproduce:
1.
2.
3.

Comment 1

[Anne (:annevk)]

Richard, have you planned for which build you are going to make this chart? Do
you have time to update it? I think you'd best start with the CSS 2.1
specification. After that one is covered, for let's say, Firefox 0.10, you could
add other properties, like ':-moz-first-node' or something similar.

Comment 2

[Richard York]

I'd be happy to write and maintain the chart.

Ideally to help with maintainance, I'd write an app in PHP or Perl and make the
chart as "self-maintaining" as possible so that Moz devs and volunteers could
update as new builds roll out with new features. I already monitor the
CHANGELOG, so it wouldn't be a big deal for me to do this either.

I was already thinking of doing a self-updating support chart for my own website
(http://www.smilingsouls.net) to go along with the release of my Beginning CSS book.

As far as which build to start with.. I think the focus should initially be on
Firefox 1.0, and of course CSS 2.1 but I'd be quickly getting CSS 3 CRs in
there, people need to see how far ahead Mozilla is when compared to Explorer. :-)

I'm finishing my Beginning CSS book ATM, so I could start putting this together
in about three weeks.

Comment 3

[Anne (:annevk)]

That would be great. I believe Perl is the preferred language around here, so if
that isn't a problem, I think you can start ;-)

Comment 4

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

What about those of us who think support charts are a bad idea?

Comment 5

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

While I have mixed feelings about support charts, there are many -moz-* features
that we definitely don't want to advertize, and something like this *definitely*
has to be run by Gecko developers before becoming public.

Comment 6

[Boris Zbarsky [:bzbarsky]]

How are we defining "support", exactly?

Comment 7

[Richard York]

I was thinking of a comprehensive documentation of what is and isn't supported
down to the very last property, keyword, at rule, etc with the traditional
yes/no/partial/buggy.. etc notation accompanied by comprehensive test suites
that show the feature in action er not in action whatever the case may be. The
CSS specification a feature is currently in conformance with. But a back-end
application abstracted enough that it could be applied to markup, ECMAScript,
etc, should documentation in those areas become necessary.

Anyone can do a simple yes/no chart, like the one the Safari team has up, but
that's not as useful to the average developer, IMO. There could be links with
predefined queries to search bugzilla for related bugs. Suggested work-arounds
for WON'T FIX, etc. Documentation like this would be appreciated by the web
development community and I think even expected by the open source community.

If you'd rather people didn't know about the existence of certain -moz-*
features, that seems somewhat counter-intuitive to the open source process to me
(but counter-intuitive to open standards, I realize). I understand the quandary
there. On the other hand, it doesn't really matter to me personally whether
these are documented or not, just that a more comprehensive guide of what is
implemented in Mozilla is available. I'm more interested in open standards. I've
done my share of searching through bugzilla to get a feel for what's available
and what areas of development the devs are focused on and it isn't the easiest
way to track down information.

I wouldn't mind putting in access restrictions to whatever you don't want made
public.

I suppose, just outline what you want me to provide and I'll build the  
application. I'd be happy to see anything at all put up.

Comment 8

[Anne (:annevk)]

I guess the problem with -moz- properties/selectors is that people may tend to
like them and make pages depend om them and eventually we can't remove support
for it anymore.

Comment 9

[Boris Zbarsky [:bzbarsky]]

(In reply to comment #7)
> I was thinking of a comprehensive documentation of what is and isn't supported
> down to the very last property

The problem is how one defines "support" for a property....

I guess it feels to me, more often than not, that support charts simply gloss
over known issues.

> accompanied by comprehensive test suites that show the feature in action

_That_ would be a great step up over most support charts.

If done right, we could get Opera (and maybe Safari) people to pitch in on said
test suites....

Creation of such a test suite would be useful no matter what.  If nothing else,
it can be used as part of our regression tests and maybe contribute to the CSS
WG test suites.

> If you'd rather people didn't know about the existence of certain -moz-*
> features, that seems somewhat counter-intuitive to the open source process to
> me

Well, some -moz-* features are really meant for "internal" Mozilla use only.  We
don't want web content relying on them because we want to reserve the right to
change them with no notice...  We're not trying to hide them, just trying not to
advertise them to people who will misunderstand the idea.

Listing them as such ("subject to change without notice") may be a decent approach.

Comment 10

[fantasai]

> maybe contribute to the CSS WG test suites.
Hixie-conformant test cases (i.e. CSS conformance tests) are only suitable for
demonstrating, quantitavely, how broken a feature is. They're horrible as far as
demonstration goes: a series of green boxes isn't enlightening to the average
web developer. :)

Comment 11

[Richard York]

> I guess it feels to me, more often than not, that support charts simply gloss
> over known issues.

Right, I agree, I like a more blog/wiki style approach something community
maintained in the spirit of Gecko itself, something that could serve as a manual
and a testing tool too. Personally I like the ones created by Peter-Paul Koch at
quirksmode.org. The web already has a rather **** selection of support charts,
why build another **** chart? It needs to stand out and be worthy of the Mozilla
name. Eric Meyer's mastergrid was such a chart, but it's not up-to-date anymore.

> Well, some -moz-* features are really meant for "internal" Mozilla use only. 
> We don't want web content relying on them because we want to reserve the right 
> to change them with no notice...  We're not trying to hide them, just trying 
> not to advertise them to people who will misunderstand the idea.
>
> Listing them as such ("subject to change without notice") may be a decent 
> approach.

Right, I understand that. Best not to create a problem with them. Of course if
you put up a disclaimer saying these features can change without notice you're
pretty much justified, IMO anyway, in changing them if and when the time comes.
I certainly see why you don't want to encourage their use though.

The test suites may not be interesting to most developers, but that can allow a
resource like this to have more than one purpose. Be a more user-friendly doc
portal and somewhere the devs can test and keep track of things as well. I think
a portal like this belongs on a browser website though. Microsoft offers test
suites in their CSS docs page, but these aren't very good and are often driven
by script instead of being pure CSS test suites.

Comment 12

[Anne (:annevk)]

I have been doing some work that related to this: bug 281960. (CSS support
chart, without the test cases, et cetera.)

Comment 13

[Richard York]

(In reply to comment #12)
> I have been doing some work that related to this: bug 281960. (CSS support
> chart, without the test cases, et cetera.)

I'm available to help, if you need it. I haven't been able to get a start on
this due to other responsibilities. Just let me know what you need.

Comment 14

[Reed Loden [:reed]]

--> MDC :: Doc Request