2005-08 Archive

Writing

In many of my pursuits up to this point, my main writing has been in mailing lists or blogs, comments or wiki edits. I have never put together something long enough to be considered for an actual "article": I've never had the reason to, since it seems that so many things are covered better by other writers.

However, lately I have found that this is no longer the case: I have a number of topics on which there is no accurate, thorough documentation. Specifically, some topics related to RDF have general, shorthand notation for people familiar with the topic, or the APIs, or wishing to become familiar with a particular API, but no general documentation for them, or more general documentation for how to work with RDF.

I'm trying to keep this relatively quiet, but I've been working on changing that. A few people have seen perliminary efforts of mine to write a tutorial on how to create an application based on RDF. I've not seen anything else that is geared specifically towards application development, so I think that this may be an article which has some relatively wide reaching potential. However, I'm not sure exactly what I plan to do with it once I finish.

Typically, it seems, these kind of articles are used or written for sites like XML.com, or IBM developerWorks. (My goal is to write a series of articles similar in depth to those, for a general idea of my target.) However, although in the past these articles have tended to be based around RDF/XML: meaning that they do, at least somewhat, related to xml.com's main thrust - mine are doing their best to ignore RDF syntax altogether. With that in mind, I'd be hard pressed to even advise XML.com editors to take these articles into consideration for their site.

It's too developer oriented for LinuxJournal or Make Magazine. It's too little XML for xml.com. It's too hardcore for anything from o'reilly that I'm aware of. This is leading me to think that I'm just going to have to self-publish them. With no reputation to speak of, I don't have a lot of clout with anyone out there, and clout or reputation are both needed in order to get much of anything done in the publishing world so far as I'm aware.

I'd be interested in feedback anyone has on the matter of course: if you want a sneak peek, talk to me and we can talk about it, and any ideas you might have as to where something would go. I'm relatively inexperienced in the whole field of writing-for-money (or for other people), so I may be missing something obvious here.

Feedback, as always, to crschmidt on irc.freenode.net, or crschmidt@crschmidt.net.

Julie

Julie now supports ASK SPARQL queries.

15:20:20 < crschmidt> ^q ASK { <http://crschmidt.net/foaf.rdf#crschmidt> 
                         foaf:nick ?n. }
15:20:20 < julie>  True
15:20:22 < crschmidt> ^q ASK { <http://crschmidt.net/foaf.rdf#crschmidt> 
                     foaf:nicka ?n. }
15:20:22 < julie>  False

Code:

if res.is_bindings():
    # oldcode
    pass
elif res.is_boolean():
    return res.get_boolean()

Julie Now Supports SPARQL Construct

Construct queries create a stream of statements as triples, which are returned, up to 900 characters. Not really much use for this one on IRC.

However, I'm just being complete.

Noets Addition to Planet Swhack

In an effort to stamp out the rising tyrant, "What Planet Is This", noets is now being added to Planet Swhack. Please feel free to use this tool for any reason you deem neccesary, either via PM, in #swhack, or in one of the other channels blogbot hangs out in (#julie, #d8uv.com).

Thank you for your participation.

Julie: SPARQL CONSTRUCT as rules

As discussed in SPARQL CONSTRUCT as Rules, julie now has a way to persist the output from SPARQL CONSTRUCT queries, as well as an improvement to the CONSTRUCT query output.

When passing a CONSTRUCT query to julie, she will now tell you the number of triples returned, and the first three triples in the standard Redland representation. If you wish to persist these triples, you can use the ^construct shortcut: ^construct { ?a ?b ?c. } WHERE { ?a ?b ?c. } As always, namespaces are prefixed automatically.

If you have feedback on this implementation, or thoughts on it, feel free to contact me, or stop by #julie on irc.freenode.net to pass on criticisms or comments.