Lists Home |
Date Index |
----- Original Message -----
From: "Michael Brennan" <Michael_Brennan@Allegis.com>
> > From: Paul T [mailto:firstname.lastname@example.org]
> > This wins with no question, I think. And I think
> > the buzzword should be not RDF, but RDDL
> > ( not the current version, sorry ;-( ).
> I agree. I'm curious what you would like to see changed in RDDL, though.
1. The process.
I'm kinda tired of W3C geniuses silently and suddenly
"Leading the Web to its Full Potential...".
Too much politics around RDDL.
2. In my oppinion, current RDDL is a scientific
stuff that can not be used to solve a 'real problem'
RDDL makes some step into 'right direction',
but I don't get many of 'design ideas' behind
current RDDL ( yet another declarative
XML-ish 'language' )
I guess very few developers have enough
time for it.
4. A real problem is "what should I do when my
software encounteres the 'unknown tag'". It can
be solved on namespaces ( like RDDL does ).
However, in my oppinion, it is better to be
solved with a single and trivial attribute
rddl-hook="URI" or something.
5. The real challenge is on the server side.I even
designed some funny mix of DNS and FidoNet to
write a 'universal distributed DNS-alike resolvers'.
Unfortunately, to implement that stuff it would take,
like, at least, two-three weeks. I'm not doing that
a. Nobody is gonna pay for that.
I think that the desire for 'web services' is mostly
mythical and there would be many funny things with
the first attempts of implementing the 'webservices' in
their full glory. ( also, see 'PS' section )
b. Even I can implement it, I'm having a hard time
to find some other people, who may be interested
in *using* such a distributed 'DNS' ( or RDDL ).
Maybe people don't really like distributing
I hardly understand how to build an *application*
for that 'RDDL'.
If there is not a big number of people who need
the distributed *applications* *right now* -
why bother with that RDDL at all?
Just place some human-readable documentation
at the end of namespace URL and that
would work. ;-)
PS. As to 'webservices'. I have messy feelings.
I have two experiences using the 'webservices'.
1. DTDGenerator Frontend on www.pault.com.
I just wrapped Michael Kay's perfect DTDGenerator
with perl ( and then python) script - many people
use that 'service' it every day.
The *only* problem with that 'webservice' was to
maintain the website up and running 7/24/365, but
because it is on Linux - the uptime of 100 days is
normal. No XML involved at all.
2. ( True story ) I have missed a job interview, the
first week I installed Windows XP on my home
computer. I have bought Windows XP professional
pre-installed on a *perfect* Dell CPU, to see
how robust is it on 'ideal' hardware. No comments
on overall reliability, comparing to Win 98 - that
would be irrelevant.
Now back to the point.
XP was using the 'time webservice' to synchronize
the system clock. Because of bugs in XP or maybe
bugs on the 'service side' - I would never know -
my nice computer was occasionally, but constantly,
crushing my system clock in subtile ways
( usually +/- one hour ).
After some magic dance (as it always happens,
when trying to configure some MS product), I've
disabled that 'time syncronization' ) so I now (finally)
can trust the time displayed in lower right corner
on my desktop.
But I *did* missed an important job interview that
I've been scheduing for one month and I will never
forget this ... experience ...
Sometimes, when I read that some hospital is now
building on .NET
- I wanna cry.
> Is it time to revisit RDDL?
I think that it always be the right time.
The problems are still there, they just
decided not to touch it. Politics, politics.