php interface

Daniel O'Donnell writes:
 > As far as I can tell--I'm a novice PHPer but experienced XML/XSLT
 > person, so likely biased--the way to go for importing z3950 put into
 > refdb is via XML: yaz provides good MARC to MARCXML functionality, and
 > writing a stylesheet that mimics marc2ris probably wouldn't be to hard
 > if the target were risx.
 > 

I don't want to talk you out of this, but in the end the data have to
be piped into refdbc anyway. If you convert MARC data internally to
MARCXML and then to risx, you basically duplicate the existing MARC
to RIS conversion. This at best duplicates the efforts to get better
MARC support, but at worst it causes inconsistencies: The SQL data
should not depend on the path how you import the data (command line
vs. web interface). I'm not generally opposed to going through
MARCXML, but if you follow this path, you should provide modular code
which can be used from the command line as well.

 > The advice on the net seems to be to avoid mixing perl and php; since
 > xml handling is built into php 5, this seems both a safer and a more
 > elegant approach.

We're not talking about mixing Perl and PHP. The PHP code will have to
pipe data into refdbc (unless someone codes the client networking code
directly in PHP). PHP will open a shell for this purpose, and PHP does
not even notice whether the first command in the pipe is a C program
(refdbc) or a Perl script (marc2ris).

 > 
 > B. 
 > As part of my experimentation, I thought I'd also play a bit with the
 > refdb_php. In particular, I'd like to do two things:
 > 
 > 1) xhtml-ize the current html (make it well formed, use CSS instead of
 > tables for positioning)
 > 

You should keep in mind that the original web design (which is
replicated in the current PHP interface) was created by a person not
very capable of designing user interfaces (I'm talking about
myself). If you start fiddling with the HTML/XHTML code, you should
feel free to do a redesign as well, if it helps to improve your
code. I'd appreciate clean XHTML + CSS code.

 > 2) change the output to either risx or tei xml (more possibilities for
 > display and exchange)
 > 

You should probably provide several output formats (see the Emacs
interface). For entering new references we should think about
designing an interface that does not directly show the underlying data
format.

 > Any thoughts on either? I know nothing about making packages, so I'll be
 > doing all this on my local installation as a hobby. I'm not quite sure
 > how one posts progress or enables collaboration but I'm happy to do so.
 > 

The easiest way is to sign up as a RefDB developer and use the
subversion repository to publish your work and to collaborate. Let me
know if you are interested.

regards,
Markus

-- 
Markus Hoenicka
[email protected]
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de


Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642