OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   Re: [xml-dev] RDDL (was RE: [xml-dev] Negotiate Out The Noise)

[ Lists Home | Date Index | Thread Index ]


----- Original Message ----- 
From: "Bullard, Claude L (Len)" <clbullar@ingr.com>


> bomb, he was inventing the future.  To 
> me, RDDL looks like the linkbases of old. 
> Big ol'databases were built that way in 
> the Sixties and they worked.  I'm not 
> anti-RDDL and I don't think you are. 

I am not. I'm also not 'against' the COBOL
or 'against' anything else. BTW, that was 
stupid me who first wrote that 'RDDL' 
word, saying that it 'looks closer, than RDF'.

> I'm not fond of pine tree solutions: 
> they kill anything that tries to grow 
> around them and leave sticky useless 
> cones on the ground.  So RDDL is fine 
> until it becomes an excuse to dismiss 
> alternatives.  That is the politics of 
> some supporters, but not the design of RDDL.
> Spy Vs Spy.

Agree.
 
> >'there should be some way
> >to share the 'semantics' of 'some tag' 
> >( or 'namespace' ) on the web'
> 
> On the Web, or among some set of applications 
> or information owners that use the Internet 
> as a hosting and transport medium?

Agree, this is 'better'. But I think that 
'worse is better' implies 'on the Web'.
 
> >And also RDDL is right that URLs 
> >should be involved ( how else can it be, 
> >it is the Web ;-) 
> 
> I won't fall for that bait.  ;-)

I do fall. I think that the less entities - 
the better. No to 'URI'. URL only.
'Universal link'. Not ideal one, 
but we have it everywhere already.

> >But there is a lot of other problems 
> >if thinking about this possible 'semantics linking'. 
> >Caching, certification, distribution, e t.c.
> >I've spent one year thinking and I think that 
> >it would take years to get it right. Simple 
> >things are complex .
> 
> Errr... at the risk of immolation, .Net and 
> it's competitors are reaching for that brass ring.

And the application they have in mind is?

I remember that their idea was that my 
car would be able to buy a ticket to 
opera or something. Like I trust my old 
honda to know opera better, than I do.

Java (when it was called Oak) had pretty 
clear task to solve and they did it right. 

That's also what makes  Java sucessfull in 
current domains - the core was healthy 
because it was solving the particular task. 

> >And ( most important ) all this stuff would not 
> >fly until somebody would try implementing a 
> >real-life project with that. 
 
> See above.  God help them if they are wrong because 
> they are betting their companies on it.

... OpenSource *development platforms* 
are powering  *way* too many companies 
these days.

I remember how it was 15 years ago. 
There was almost *no way* to make 
money if  using opensource language. 
( and there were very few of them ).

Compare it to the current situation? 

The difference is *dramatic*.  I know 
some young developers who have 
not touched a *single* commercial 
tool in a very long time!

The Internet is doing the same thing 
to big Software Co's that Apple II
did to IBM. *Exactly* the same thing.
IBM is alive, but I remember reading 
some pretty scary stuff about the 
restructuring they had to do.

> >The 'phonebook' was a 'killer application'
> >for the web. What is the 'killer application'
> >for RDDL? 
 
> I dunno.  

Too bad. Then we need to wait for next 
TBL, because I also don't know what it 
may look like ... 

For example, SPAM filter that I've written,
should scale for millions of users, sharing 
the knowledge about the SPAM ;-) so that 
would make a huge distributed SPAM filtering 
network ... But I found that people don't like 
the idea that *any* information from their 
computer will go elsewhere, so it does not 
work as I wanted it to work, actually...

There is distributed.net also .. P2P ... napster,
ICQ ... livejournal.com ... some pieces here 
and there ... but they look too exotic ... 
phonebook is what any person may need ;-)

> Maybe the idea that a language
> needs a killer application is inherently another 
> magic spell.  

Well, I have not seen and have not read of 
a different pattern.

All the sucessfull projects that I know 
are a result of some person solving the 
particular task. I don't know of any exception.
( It does not mean that the task would 
*persist*, for example, Perl has been 
created to write some system, but when 
writing Perl , Larry Wall got so excited 
about perl, that the system itself never been 
written ;-) But still. In the beginning there always 
is a Problem, I think.

> It could be that RDDL does 
> some things that a lot of people need and does 
> it in a way that is easy to understand and 
> apply.  

If there is a lot of people, who need RDDL, 
then we should get a particular 
application idea from them pretty fast.  

We'l see it immediately, if somebody 
really needs some distributed 'semantic web'. 
I doubt that, but I'd be *very* happy to 
be wrong.

> Maybe it is an Angel application 
> that tries to preserve life by giving others 
> time to live it instead of trying to implement 
> grotesque baroque code.  Don't Play Bach If 
> They Came To Boogie.  I'm waiting to hear what 
> that something is.

So am I. 

> >Some *particular* task.
> 
> Always better to have a problem to solve. 

I belive it is unavoidable in software 
development. I think this is what makes
software development different from math. 

Those, who claim that software 
development === math would never 
get a decent salary as a software developers.
Of course Math helps a *lot*, but it does 
not equals to software development. 

> <rant>Those of us who 
> wanted SGML On The Web did so because we knew 
> from experience precisely where HTML would 
> run out of steam.  We had to steal ideas, change 
> the name, and virtually knife a community to 
> do that even though many of us were nursed 
> by that community.  Ask yourself where we would be today 
> if Dr. Charles Goldfarb had really fought XML.  
> Remember, he put his entire life, career, and 
> fortune into making markup a workable system. 
> It takes a helluva man to get it that far, then 
> hand it over to kids for the last two inches. 
> But when you look at SGML, it was there to solve 
> a bigger job and where it succeeded or failed, 
> we who came behind him found the border cases. 
> That made XML a lot easier to spec and made 
> some people look like geniuses.  No one denies 
> it now, but we built on the shoulders of a man 
> who did the dirty work before we got here.</rant>

Happens all the time. Nobody remembers that 
it was RMS who have actually written Linux.

I belive that you are very much right and I 
feel sorry for some brilliant guys who've 
done a hard work with SGML and I also 
feel that I got brainwashed by some XML 
propaganda, not paying enough attantion 
to the roots. The only question 

> Ok.  I leave it to the RDDL designers to answer 
> that one.  As I recall, they had definite applications 
> in mind, but it was designed over the Christmas 
> holidays and I was playing guitar in my new 
> house last year.  I missed it, darnit.

Yeh, that year was really special. I wish 
this one will be better, at least XML-dev 
was down this time, so the XQuery 
stuff did not explode ;-)
 
> >I'm sure there was some particular application 
> >for markup. 
> 
> Circa late Sixties:  publishing houses needed a 
> way to move manuscripts among dissimilar and 
> non-interoperable printing systems, thus emerged 
> GenCoding.  IBM needed a way to unify their contract 
> and other legal document work.  Thus emerged GML. 
> Scribe was borrowed from for that, but originally, 
> lawyers needed a better way to do their work.  
> Charles wrote a history of that and it is at his 
> web page.

He-he. As I was suspecting. Printers. Somebody 
just wanna 'print that damn document'. There is 
always some simple problem to be solved. It is 
just not easy to know what it was, if the person 
who managed the solution forgets about it
( happens all the time as well )

> >In the healthy situation, problem *always* comes 
> >*before* the solution. 
> 
> Engineering over art.

Yes. 

Engineering over Math.
Engineering over Art.

Software development is not about making 
pleasure to users of software products! 

The last thing I want is my word processor 
start singing nice songs.  *Nobody* wants that.
Software development is to turn computer into slave.
No art allowed.

Art is for people.

Rgds.Paul.






 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS