[dev] Re: [Bug 1052] Changed - multilingual support in IMP is broken/misdesigned

[Michael M Slusarz]

[Posting to dev list]

Quoting bugs at bugs.horde.org:

| http://bugs.horde.org/show_bug.cgi?id=1052
|
| *** shadow/1052	Mon Jun 23 16:08:35 2003
| --- shadow/1052.tmp.12651	Mon Jun 23 22:53:51 2003
| ***************
| *** 497,499 ****
| --- 497,527 ----
|   report will be marked as closed.  If you would like to see these
| requests
|   implemented, it would be best to subscribe to either
| il8n at lists.horde.org,
|   dev at lists.horde.org, or imp at lists.horde.org.
| +
| + ------- Additional Comments From jshin at jtan.com  06/23/03 22:53 -------
| + If you insist that bugzilla is only for genuine *bugs* and is not for
| feature
| + request/enhancements (FYI, 'bug' in Mozilla's bugzilla and other
| bugzilla  -
| + Gnome, FontConfig, RedHat Linux - is NOT limited to literal bugs), this
| is
| + invalid as you marked. However, is it really the case? At the top of
| this page,
| + I can find:
| +
| + This is Bugzilla for tracking bugs/features/requests/etc for the Horde
| Project
| +
| + According to the above, 'bug' here rather means 'issues' (not limited
| to
| + malfunctioning), doesn't it? Moreover, why does this bugzilla have
| 'enhancement'
| + as one of values for 'severity' field if it's only for 'genuine bugs'?
| +
| + If all the above are just remnants of the outdated bugzilla
| configuration and
| + horde bugzilla is NOT for general issue tracking,  that's fine and I'm
| willing
| + to discuss/tackle the isseu at horde-i18n  list. However, it'd be
| better to
| + update the bugzilla configuration not to confuse people like me.

I agree with you that the wording tends to point this way.  But there are at
least 2 reasons why this is so.  1) the system is old (relatively speaking,
at least in "internet" time) and 2) at last check, the maintainer of the
bugs server could not be reached at all.

However, it is pretty much the case (at least for the last year IMHO) that
the horde bugzilla system has been limited to bug reports only.  Feature
requests, etc. should NOT be in the system mainly because noone else will
see it - unless the report is assigned to you, other developers are not
going to see it.  I, for one, have not looked at reports assigned to others
in a good long time (1+ year) so if there are featrue requests in those
reports, I sure am unable to give feedback.  I am sure the other developers
would concur in this statement.

Once the whups system is up to replace bugzilla, hopefully this confusion
will be remediated.

But, that being said, I personally think my previous note made very clear
both 1) that our bugzilla system was no longer being used for these kind of
requests and 2) clear identification of other places you could bring the
requests.  As mentioned, we would love to improve the internationalization
support of IMP - the mailing lists I mentioned would be the place to bring
this though (I, personally, have little/no experience with language
programming since I am a monolingual idiot. I was only assigned the report
due to the MIME aspects of it, most of which have been addressed in some
form by the code in HEAD.  Posting to these lists would no doubt result in
much more active participation than being stuck in my bug list).

| +
| + > Additionally, I personally have not seen RFC 2331 encoded
| + > messages ever, so
| +
| +   E-mail bodies are NOT encoded in RFC 2231. RFC 2231/2184 are for
| parameters of
| + email header fields.

You misread what I wrote.  I have never seen a message with RFC 2231
encoding - with the term 'message' obviously including the headers portion
of this message.  I did not distinguish between 'body' or 'headers'
anywhere.

| And, they're generated by standard-compliant MUAs
| like
| + Pine(http://www.washington.edu/pine)  and stadnard-compliant web
| applications
| + (written in Perl, PHP, Java, VBscript, etc) everyday (see
| + http://bugzilla.mozilla.org/show_bug.cgi?id=162765)

I didn't say they don't exist.  And I definitely didn't say we don't want to
support it.  It is just that I have *personally* never come across a RFC
2231 encoded header.  And I can quite honestly say that I have viewed
thousands, if not tens of thousands, of MIME headers in the past few years.
 And I am the person who generally fixes the Horde/IMP MIME code if a
message doesn't render correctly, so you would figure that someone out
there in the IMP community would have discovered this problem and brought
our attention to it.  Aside from you, there has been noone; thus, I can
safely say _at this time_ that it must not a critical problem.

Once again, it would be great to eventually have this support.  And any
patches/code that could be applied to HEAD would be fantastic.  But does
the lack of this support make Horde/IMP multilingual support
"broken/misdesigned"?  Hardly.  Name another webmail program (open source
especially) that supports multilingual support better.  Squirrelmail seems
to be another popular webmail choice, but having plugged around with it a
bit it is *nowhere* near as good as the code in HEAD (which I personally
have done none of - accolades should go to the other developers,
particularly Jan Schneider).

Obviously, we can (and will) continue to make multilingual support better. 
But this is not a bug report but a feature request and therefore is why I
closed the ticket.

Finally, this message is not meant as some kind of 'crusade for Horde world
dominance/we will crush all in its wake' statement.  Rather, several of the
issues you brought up are fairly important, and seemed like worthy
candidates for comment at this time.  Thanks for helping us make Horde/IMP
a better product.

michael

[Here's my lawyerly disclaimer: All views expressed in this message are my
own and do not reflect on other Horde developers. There, how was that? :)]

______________________________________________
Michael Slusarz [slusarz at bigworm.colorado.edu]
The University of Colorado at Boulder