Hi, I just wanted to ask whether any discussion was held on using xml for developing koha? www.w3.org says it is good to separate the data from the way it looks (data from html). Do you think that future direction of koha development could benefit from it (like language translations)? Benedict
On 2003-10-03 15:56:15 +0100 Benedykt <kb2qzv@poczta.wp.pl> wrote:
I just wanted to ask whether any discussion was held on using xml for developing koha? www.w3.org says it is good to separate the data from the way it looks (data from html). Do you think that future direction of koha development could benefit from it (like language translations)?
To me, there are three sides to this. First, using some xml protocols in koha: a good idea in general, but comes with some nasty research, design and robustness problems if we want to accept XML data. I think XML output is sure to start appearing in 2.2 (RSS is already out there somewhere) and not spread to any core use until 2.4 or later, if at all. We may be forced this way if users need the "Z39.50 in xml" stuff or the OAI protocol. Second, using xhtml output and CSS layout: makes display much cleaner and easier, IMO, and more accessible if done carefully. A really good idea and, personally, I hope that most of the code work can appear in 2.2, but it depends on many other people too. Finally, using xml display technologies: xml, xslt, etc. A nightmare IMO. Don't even go there with koha. I've been there before and am not keen to go back. Full XSLT is apparently turing-complete and it seems really dangerous not being able to tell a template from an infinite loop. If people really want to do this, let them use XML output from koha and reformat it. -- MJR/slef My Opinion Only and possibly not of any group I know. http://mjr.towers.org.uk/ gopher://g.towers.org.uk/ slef@jabber.at Creative copyleft computing services via http://www.ttllp.co.uk/
MJ Ray wrote:
Finally, using xml display technologies: xml, xslt, etc. A nightmare IMO. Don't even go there with koha. I've been there before and am not keen to go back. Full XSLT is apparently turing-complete and it seems really dangerous not being able to tell a template from an infinite loop. If people really want to do this, let them use XML output from koha and reformat it.
iirc, chris suggested a Perl XSLT engine when we had to choose our tools for 2.0.0 He thought it was a "heavy idea", so was not very enthousiastic. I tried to install the tool, but never succeded. The conclusion was : - it's a complex solution for developpers - it's a nightmare solution for users => it's not our solution :-) -- Paul POULAIN Consultant indépendant en logiciels libres responsable francophone de koha (SIGB libre http://www.koha-fr.org)
On Mon, Oct 06, 2003 at 10:59:38AM +0200, paul POULAIN said:
MJ Ray wrote:
Finally, using xml display technologies: xml, xslt, etc. A nightmare IMO. Don't even go there with koha. I've been there before and am not keen to go back. Full XSLT is apparently turing-complete and it seems really dangerous not being able to tell a template from an infinite loop. If people really want to do this, let them use XML output from koha and reformat it.
iirc, chris suggested a Perl XSLT engine when we had to choose our tools for 2.0.0 He thought it was a "heavy idea", so was not very enthousiastic. I tried to install the tool, but never succeded. The conclusion was : - it's a complex solution for developpers - it's a nightmare solution for users => it's not our solution :-)
Yeah, We expiremented with Axkit, and while impressive, it did cause a significant performance hit. And did tie us to mod_perl and Apache. Where html::template (which of course the templates dont have to be html .. they can be anything you like (including XML)) provided a more portable solution. Chris -- Chris Cormack Programmer 027 4500 789 Katipo Communications Ltd chris@katipo.co.nz www.katipo.co.nz
participants (4)
-
Benedykt -
Chris Cormack -
MJ Ray -
paul POULAIN