[pmwiki-users] (:dataset:) directive/concept

[Henrik Bechmann]

Patrick,

I'm looking for some strategic advice on my next main issue in using 
PmWiki as a simple content management system (so far so good by the way, 
my ability to create (:includefieldpage :) to support a Farm of fields 
cross-referencing each other rounded out the basic requirements).

The next main issue is picture management, as the website that I'm 
Webmastering is essentially in a kind of magazine format, with by now 
(I'm guessing) thousands of pictures. Nonetheless the main problem is 
just being able to see what's there, so that authors can reference 
existing pictures in their articles. Nothing too fancy required or 
wanted (all the tools I've seen so far are way overkill). PmWiki already 
has a competent upload facility, and FileZilla can be used for bulk 
uploads, and IrfanView for bulk processing.

I can see fairly easily listing picture repositories (perhaps with a 
PmWiki directive extension of some kind, linked to InterMap listings), 
and from those automatically generating pages of thumbnails (based on 
images found on disk in the repositories), included generating standard 
file caches of the thumbnails, and having a next/previous page listings etc.

What's missing is the ability to associate titles, descriptions, and 
keywords with pictures, and I don't see that being handled in the 
automated way I would like with the existing recipes in the cookbook 
section.

So in other words I need a simple, small-scale, text database with which 
I can associate picture attributes (titles etc) through the picture 
name. I also need for the authors to be able to update this database 
fairly readily.

I'm thinking that most of the pieces are available in the PmWiki 
environment to solve this problem: php image generation, the PmWiki 
directive engine, your new Input Forms. But the piece I don't see is the 
repository for the simple text database. And I wondered if you could 
give me some strategic direction for that.

I was thinking, for example, about a (:dataset name=DataSetName 
fields=FieldA, FieldB, FieldC=image, FieldD, FieldE=wikipage key=FieldA 
actions=insert,edit,delete displaymode=table:) directive which would be 
analagous to the (:table:) directives, inasmuch as it would have nested 
parts, something like:

(:dataset ... :)
(:fieldset:)
(:field FieldA:)
(:field FieldB:)
...
(:fieldset:)
...
(:datasetend:)

Something like that.

I envision that such a structure could have fieldset additions inserted 
by script, so that I could generate an automatic inventory with blank 
fieldsets for newly uploaded images, and I might even be able to write a 
popup entry form per fieldset, with some Javascript (is this a no-no in 
PmWikiLand?) to close the popup and refresh the underlying browser list 
for the user.

One of the benefits of keeping all this inside the PmWiki environment is 
that the keywords etc. would be searchable.

So the question: Is this a strategically sound direction? Is there 
another direction I should consider (for example your entry for the 
$WikiDir variable makes intriguing mention of the possibility of using 
it for  "something more advanced such as ... a .dbm-file")?.

Any thoughts you may be able to share about this would, as always, be 
most appreciated.

Best,

- Henrik

-- 

Henrik Bechmann
www.osscommons.ca
www.bechmannsoftware.com
Webmaster, www.dufferinpark.ca