On Nov 15, 2007, at 6:55 AM, Thomas Bruederli wrote:
You are right, the specification isn't very precise. I've just added a "tested browsers" section.
I took a look at it, I think the change makes what to use more clear .
It's a wiki and registered users are able to add the browsers they successfully tested with.
I added a "Browsers Reported to work" section, then added a browser
known to work here.
On Nov 15, 2007, at 8:05 AM, markevich@inbox.lv wrote:
I was using Firefox 2.x (Windows, Linux), Opera 9.x (Windows,
Linux), Internet Explorer.
You might want to add the version of Opera you use to the "Browsers
Reported to work" section.
I don't think that browser names should be displayed in the
requirements, because there are a lot of differences in the
browsers engine realization (even in the same family).
Exactly my point. Since the implementation details of HTML, CSS, and
ECMAScript are different between browsers and browser versions, the
existing requirements indicated that many browsers should work when
in reality they might not.
The key here is to specify generic functionality that is used in
the Roundcube that at least must be supported in order to work.
Yes, requirement specification is a very difficult task for _any_
piece of software. You don't want to be too detailed and scare people
away, or make the requirements difficult to read or understand. On
the other hand, there is a limited amount of testing done. It would
be a very large task to find and test all browsers that fit the
generic requirements. It is my opinion that those that encounter
problems with RoundCube and use a browser not listed under "tested"
are savvy enough to look for a requirements page, see their browser
of choice is not listed, and figure out that their problem is likely
a browser related issue. That can help the developers with better or
more specific bug reports. The people that use any of the browsers
listed likely won't read the requirements page ;)
If we’ll add some browser’s names to the list it would mean that
we’ll give support obligations to the community to 100% support for
these browsers and it wouldn’t be right.
If the developers test their work using the browsers they list, isn't
that true ? Would a developer test in a popular browser and not fix
an issue discovered ?
My advice is to keep things like they are now – if your browser is
supporting specified functions in the requirement it should work,
if it’s supporting and not working – create ticket and we’ll try to
investigate the problem.
The specification of "Javascript" is very general. There are several
versions of JavaScript, and technically the latest versions are
correctly called ECMAScript. Although the CSS2 requirement excludes
older browsers. For example Netscape 3.0 supported "Javascript" and
yet that browser isn't likely to work ( nor is it worth my time to
dig out an installer to check ). A browser may claim to support
"CSS2" but not support the specific features used by RoundCube.
When browsers support ECMAScript 4, there may be features RoundCube
would like to take advantage of. If that happens, only specific
browsers and browser versions will work.
Many people don't know what any of those things listed under "Browser
Requirements" are, but _do_ know the name of the browser they use.
Some however don't even know the name of the browser, but that is a
different topic ;)
Charles Dostale System Admin - Silver Oaks Communications http://www.silveroaks.com/ 824 17th Street, Moline IL 61265
List info: http://lists.roundcube.net/dev/