|
Extensible Programming for the 21st Century
|
|
|
I've just had my bi-monthly heart attack, when I stumble across something that
seems to scoop my entire thesis, and spend the morning frantically chasing
references to check that it doesn't. In this case, it was a fairly interesting
article (linked
to from this
SlashDot thead) about "extensible programming", wherein instead of
communicating data between components in terms of streams of characters,
á la the Unix command line, we use something a little more
structured, which at the moment translates as XML. He also brings in a lot of
together other ideas, such as Scheme hygienic macros, in a view that's spookily
similar to my own way of thinking. Worth a look.
(In case you're wondering, the thing that worried me thesis-wise was the
fourth footnote, which alludes to the fact that .NET makes translation between
source languages "almost possible" via common intermediate form. I've checked,
and I can't find anything suggesting that anyone actually does this with .NET.
If you're reading this, and know of someone who does, then please let me know.)
|
|
|
|
|
Metadata, Good and Bad
|
|
I've been looking into RDF
recently, as I'm writing an XML-based bibliography tool (by the way,
if anyone knows of an existing RDF vocabulary for describing citations
in journals, please, please, please mail me and let me know what it is),
and it looks pretty useful for a variety of things, including a
messaging/blogging/browsing thingy I've been vaguely thinking about
for a while now - more when and if I get round to writing it.
Anyway, while scouring the web for resources (there are lots, as long
as what you're interested in is RSS; I'm not), I came across a nice
antidote to
metadata. Handy to bring you down to earth if you're getting
carried away with the whole semantic web thing.
|
|
|
|
|
nXML-mode: How XML should Work
|
|
|
James Clark, writer of expat, editor of the XSLT Recommendation and all-round XML
genius, has done it again, by writing an XML editing mode for Emacs that
doesn't suck. In all fairness, PSGML-mode isn't terrible - I wouldn't have been
using it daily for the last few years if it was - but it was an SGML editor
with XML functionality bolted on, which meant it was unnecessarily
complicated, and couldn't do anything useful without a DTD. nXML-mode is still in
alpha, and doesn't have a pretty download page yet, but is nevertheless the
bee's knees, the dog's bollocks, and almost any other part of any animal you'd
care to name (except, perhaps, the dingo's kidneys). It's based on RELAX NG
schemas (and comes with XHTML, XSLT and DocBook out of the box), works fine
without any validation, supports namespaces, and autocompletes like a dream.
Hoorah!
(It may seem like I'm getting a little overexcited about Yet Another Emacs
Mode, but I'm getting my head down to write my thesis, and hence will be
writing even more XML than usual over the next few months, so I welcome this
like a plumber would welcome a new, revolutionary sort of self-assembling
pipe. Or something.)
|
|
|
|
|
|
