Hi, In article <004f01c2afa1$82eeaab0$0a00a8c0@kb2qzv>, Benedykt P. Barszcz <kb2qzv@poczta.wp.pl> wrote:

I have noticed that the french translation of the template files contains diacritical characters which are represented with the &...; sequence (I am sure it has got a name by itself).

Since I have used Kate Editor under redhat where polish diacritics are available directly I wonder whether I could leave it that way. Must I convert all the polish characters (ISO-8859-2, I hope) into the &....; html sequence?

no, just leave the Polish characters as-is. As long as the pages are correctly marked as ISO-8859-2, they should be ok. In some languages (e.g., Chinese), most characters don't even have an entity (those "&" sequences). -- Ambrose Li <a.c.li@ieee.org> http://ada.dhs.org/~acli/cmcc/ http://www.cccgt.org/ DRM is theft - We are the stakeholders

dbkliv wrote:

kb2qzv@poczta.wp.pl wrote:

...

Hi all, I have noticed that the french translation of the template files contains diacritical characters which are represented with the &...; sequence

We used the &...; sequence for want of a better solution : that way we're pretty sure the characters will display properly, but it isn't really easy to use : after a few é and ô you tend to get lost in your own translation... Any suggestion of a better solution welcomed. [...]

...

Since I have used Kate Editor under redhat where polish diacritics are available directly I wonder whether I could leave it that way. Must I convert all the polish characters (ISO-8859-2, I hope) into the &....; html sequence?

ISO-10646 (Unicode) is probably the strongest bet here.

I'd be happy to hear more about that... Nicolas Morin

Hi, the discussion about proper character's encodings is leading me to the thought that perhaps each translation team should handle it the way they want. In Poland it has been agreed nationwide that for proper web page generation one should make sure that either each page has the meta tag "http-equiv" included in the HEAD section or the web server prepends the directive every time such page is served (in its config file). But my question is what to do with those <FONT SIZE=2 face="arial, helvetica"> that can be found in "details-opac.tmpl" ? No matter what I do and whatever a meta tag I insert the FONT tag forces iso-8859-2 on my system. And I think many will experience same problem. Any suggestions Benedict ***************r-e-k-l-a-m-a************** Chcesz oszczędzić na kosztach obsługi bankowej ? mBIZNES - konto dla firm http://epieniadze.onet.pl/mbiznes

In article <20030108204355Z665253-13886+65734@ps1.test.onet.pl>, Benedict <kb2qzv@poczta.onet.pl> wrote:

But my question is what to do with those <FONT SIZE=2 face="arial, helvetica"> that can be found in "details-opac.tmpl" ?

No matter what I do and whatever a meta tag I insert the FONT tag forces iso-8859-2 on my system. And I think many will experience same problem.

I think if this is a problem, then the FONT tags should be either corrected (to use an ISO-8859-2 font), or removed. For Chinese pages, FONT FACE="arial, helvetica" tags will also force pages to ISO-8859-1 on some browsers (but not all), resulting in completely unreadable pages. Usually, if I have to fix FONT tags but there are too many of them to fix, I would replace all of them with style sheets; at least if my fix is wrong I will only have to fix one style sheet. I haven't tried this with Koha yet, so I don't know how practical this is, but I suppose it could be done. -- Ambrose Li <a.c.li@ieee.org> http://ada.dhs.org/~acli/cmcc/ http://www.cccgt.org/ DRM is theft - We are the stakeholders

In article <3E1D91B5.6090308@netscape.net>, dbkliv <dbkliv@netscape.net> wrote:

I'm curious what the problems are with representing French with Unicode (of whatever form).

I think there is no problem per se technically; in fact there are certain advantages (e.g., certain characters not available in the usual charset, e.g., "oe" in ISO-8859-1, are available in Unicode). But if everyone uses ISO-8859-1 on their computers, it would be unnecessary effort to convert everything to UTF-8 if ISO-8859-1 would be just fine. Also, I¢ don't know if people have UTF-8 editors. I¢ don't have one (actually I have one: Mozilla's Composer, although I'd prefer something more vi-like). If most people can edit UTF-8 files it may not be a problem. So my perceived "problem" is that it is not the dominant encoding. If French people have actually stopped using ISO-8859-1 and started using UTF-8, of course we should choose UTF-8.

The special French characters (é etc.) all fall into the first 256 characters of ASCII, and similarly of Unicode - at that low range, the two are nearly identical, aren't they?

Actually the ISO-8859-1 and UTF-8 codes are very different. Of course, when you are using a Unicode-capable editor, you won't know that some French characters really consist of 2 octets per character in UTF-8. -- Ambrose Li <a.c.li@ieee.org> http://ada.dhs.org/~acli/cmcc/ http://www.cccgt.org/ DRM is theft - We are the stakeholders

In article <3E1D905E.60101@netscape.net>, dbkliv <dbkliv@netscape.net> wrote:

I was suggesting ISO-10646 rather than ISO-8859-2. There would still be some work involved to convert from html entities into whatever codeset/s Koha decides to use.

Actually the work involved would depend on how much "infrastructure" you already have (or whether you want to invest some time to provide yourself with some reusable infrastructure). If you have some sort of script that can decode entities, it would be quite easy to mass-convert things. -- Ambrose Li <a.c.li@ieee.org> http://ada.dhs.org/~acli/cmcc/ http://www.cccgt.org/ DRM is theft - We are the stakeholders