Gabriel Farrell writes
> For a library school course (Projects in Digital Archives at Pratt
> Institute) I'm putting together a small online catalog for a portion
> of my music collection (20 or so records (you can read that as vinyl
> or surrogate)). I'm not worrying about an interface for adding
> records or maintenance, as I'll be doing all the backend, and am more
> concerned about having something to show the class in a few weeks. I
> plan to just keep the records as XML files (MODS or MARCXML, I haven't
> quite decided) in a directory, and serve up pages dynamically based on
> search results from swish-e. Any suggestions about general design or
> useful tools would be most appreciated.
I am not much of an expert on XSLT, but it seems that your
design is aimed to highlight techniques rather than results.
I would encourage you to think about the prolem you want to
solve and find the least convoluted way to do it.
If I were to do this project, I would use XSLT to convert
to static pages, and index those. I would not serve the
pages dynamicly because it is technically more difficult
and will generate a higher CPU time cost. I would review
the need for XSLT because I think it is not good at generating
summary pages that would allowa a user to browse. At
twenty records that is not much of a concern but IMHO a low-tech
solution that scales is better than a high-tech that does
not.
> After looking at things for a second, I realized that the "English
> Tagged View" Stylesheet is ucs-2 encoded. The "MARC Tagged View,"
> however, is utf-8, and it processes fine through 4xslt. Is there a
> reason the Library of Congress provides the "ETV" Stylesheet in
> ucs-2,
Microsoft's "save as ... unicode" ?
> Also, is there an easy way to convert it to utf-8?
You may want to try recode http://recode.progiciels-bpi.ca/.
Last time I used it, I had problems with it. But, if you
use Perl, all you do is, when you read the sheet, tell
Perl that it is in ucs-2.
>
> Finally, on the topic of client-side transformation, I've noticed that
> when I throw a
>
> <?xml-stylesheet type="text/xsl" href="http://www.loc.gov/standards/marcxml/xslt/MARC21slim2ENGLISH.xsl"?>
>
> into the example XML then MS Internet Explorer will transform it
> correctly, but in Firefox all of the metadata ends up on one line,
> with no field titles.
I doubt that Firefox does have a built-in XSLT transformation.
Most browsers don't, and they don't plan to. They are
not required to do that. Therefore I suggest you first transform
all the files to HTML. I know, it's not sexy.
You can give me a call at 718 507 1117 tonight if you want to discuss
this further.
Cheers,
Thomas Krichel mailto:krichel@...
http://openlib.org/home/krichel
RePEc:per:1965-06-05:thomas_krichel
|