Month: October 2003

Interestingly here is Ben Goodger’s original plan for Firebird web panels (via Phil Ringnalda). Confirming my suspicions, this is definitely not what we see in the current Firebird release, despite the fact that web panels are on the current feature list.

In fact, web panels are currently only even marked as different from any other bookmark by a WEB_PANEL="true" attribute in the bookmarks.html file, meaning that you can’t target it properly. Hence, when I use one of the Netscape devedge sidebars as a web panel and click a link within it, which should update the page in the sidebar, it loads that link in the main browser window, despite a target="_self" attribute.

Web panels are explicitly not sidebars, and they need to be. I’ve heard rumours that code was backed out at the last minute, but delivering no feature at all would be better than delivering just half a feature. Instead of just being left looking forward to the next release, I’m really annoyed that what I’ve been told I’ll get (and what the feature list tells me I’m getting) isn’t really there and doesn’t really work.

A new, excellent site full of everything you wanted to know about browsers, CSS and javascript by the master of all three.

Also contains the gem that IE/Mac is a Bug ridden crash prone piece of junk.

At last.

After five months of hard work I proudly present http://www.quirksmode.org

My new site contains more than 150 pages with technical information about
common and not so common browser incompatibility problems, both in CSS and
in JavaScript.

I added pages about the :not and :empty pseudo-classes, the relation of the
viewport to the HTML/CSS and JavaScript documents shown in it, the reasons
why Explorer 5.2 Mac is a bug ridden crash prone piece of junk, updated W3C
DOM compatibility tables, the reasons why Fahrner Image Replacement should
use JavaScript, and not CSS, and a short history of web development.

Finally I’m looking for JavaScript reporters. I’m trying to set up a
JavaScript news page, but I cannot possibly fill it on my own. If you’re
doing interesting JavaScript research, please write a page about it, fill
in my form and be published on my site.

Happy surfing

In amongst all these pages is an interesting Javascript alternative to the CSS-based Fahrner Image Replacement. After the recent ALA article on why we shouldn’t use the FIR (because it breaks screen readers), this is a welcome alternative, although it still needs hard testing of the kind ALA gave to the CSS version to make sure it works properly.

I’m a web developer and I test stuff locally, hence a lot of the time I’m looking at http://localhost.
I use Firebird as my default browser, which has keyword searching built into the address bar, so every time I go to localhost and my web app server is not running I get thrown to the highly classy http://www.localhost.com.au.

Fortunately this can be disabled.

Type “about:config” (without quotes) in the address bar. This will retrieve a list of all the available preferences for your browser. Type “keyword” (again without quotes) into the “filter” text box and hit <enter>.

In the returned list you should see a preference called “keyword.enabled”. Double-click this entry to bring up a box where you can change the value of the preference. Type “false” in the box and hit <enter>.

Voila keyword searching from the address bar has now been disabled (you might need to restart your browser). Things like the built in Quick Searches will still work, i.e. typing “google wellingtons” will still bring back search results about your favourite boots.

Ha! That’ll teach me to not read around the subject! Marc has already written about the problems peopleaggregator has been having in picking up who is the main “Person” in a FOAF file. Of course, I already provide dc:creator and foaf:maker properties to associate my “Person” with the FOAF document, so there shouldn’t be any problem, but as Morten and Libby point out in the comments I should also be using foaf:topic (I didn’t think this had been decided on as a method of specifying the author of an FOAF file, but there you go).

From what Marc says it looks like they’re not parsing the FOAF files with an RDF parser, but some kind of scraper to get the information out; they were looking for the first foaf:name element in the file and finding the main “Person” that way. Apart from using dc:creator, foaf:maker or foaf:topic inside the <rdf:Description rdf:about=""> element, they could use an algorithm similar to Morten’s Topic Finder which is A web service for guesstimating the primary FoaF “thing” described in an RDF/XML file.

BBC! Unified login! Now!

Jesus. I must have about ten logins for different parts of the BBC website, and I still need more. How fucking annoying is that? Can’t they have some method of sharing username/passwords across the system? I have logins for the localised message boards, blogging system (which I now can’t even find), iCan, Music meassageboards, and Christ knows what else – Celebdaq and all the other games probably all need individual logins.

To expand on my last post slightly, my actual RDF file expresses this (generated from the W3 RDF Validator):

http://philwilson.org/philfoaf.rdf#pip http://www.w3.org/2000/10/swap/pim/contact#nearestAirport genid:ARP36325
genid:ARP36325                         http://www.w3.org/1999/02/22-rdf-syntax-ns#type           http://xmlns.com/wordnet/1.6/Airport
genid:ARP36325                         http://www.daml.org/2001/10/html/airport-ont#icao         “EGCC”
genid:ARP36325                         http://www.daml.org/2001/10/html/airport-ont#iata         “MAN”

whereas my peopleaggregator.com profile expresses this:

<ns3:Person rdf:ID="pip">
  <ns5:nearestAirport rdf:nodeID="bNode1"/>
  [all my :knows statements from my RDF file here]
</ns3:Person>

<ns4:Airport rdf:nodeID="bNode1">
   <ns6:iata>MAN</ns6:iata>
   <ns6:icao>EGCC</ns6:icao>
   <ns8:acquaintanceOf rdf:nodeID="bNode12"/>
   <ns8:knowByRep rdf:nodeID="bNode10"/>
   <ns8:knowByRep rdf:nodeID="bNode11"/>
   [my manually entered peopleaggregator.com details appear here]
</ns4:Airport>

and there doesn’t seem to be any way of editing either individual triples or the file itself manually. Nor removing my current details and resetting my profile back to empty. Until I can, the emails I keep getting from Marc Canter asking me to set up relationships with people are going to be wasted effort, because whilst I can do it it’s going to establish the relationships with an airport.

I’d like to be able to talk in glowing terms about peopleaggregator.com – it has noble ideals and seems to be A Good Idea. I think they have some bugs to work on though – on first import, it thought I was Ben Hammersely (a fact which, I can assure you, as can Ben, is untrue).

vs.

And now, having cleared that up manually and having established some relationships, seems to believe that I am an airport, and that it is not I, but the airport which knows people. Most perplexing.

<ns4:Airport rdf:nodeID="bNode1">
   <ns6:iata>MAN</ns6:iata>
   <ns6:icao>EGCC</ns6:icao>
   <ns8:acquaintanceOf rdf:nodeID="bNode12"/>
   <ns8:knowByRep rdf:nodeID="bNode10"/>
   <ns8:knowByRep rdf:nodeID="bNode11"/>
   <ns3:depiction rdf:resource="http://philwilson.org/images/pip.jpg"/>
   <ns3:homepage rdf:resource="http://pipthepixie.tripod.com"/>
   <ns3:jabberID>phil@jabbernet.org</ns3:jabberID>
   <ns3:mbox_sha1sum>0a328b052de4403f6e6548d9d4389641a0cf8fdb</ns3:mbox_sha1sum>
   <ns3:name>Phil Wilson</ns3:name>
   <ns3:nick>pip</ns3:nick>
   <ns3:schoolHomepage rdf:resource="http://www.qehbristol.co.uk/"/>
   <ns3:title>Mr</ns3:title>
   <ns3:weblog rdf:resource="http://pipthepixie.tripod.com"/>
   <ns3:workInfoHomepage rdf:resource="http://philwilson.org/philfoaf.rdfJava-based web development"/>
   <ns3:workplaceHomepage rdf:resource="http://www.ibase.com/"/>
</ns4:Airport>

I’m also not sure what it’s done to my workInfoHompage property.

It’s only alpha code, but still – I thought we had working RDF parsers?