RE: different communities

["Vegt, Jan" <Jan.Vegt@softwareag.com>]

>Problem: you see problems as 'solvable'.  I see problems as 'addressable',

What's different?
Noise levels?


>but not globablly 'solvable'.

Depends on the encoding level and domain. 


>>What is fundamental?
>>Has it been solved in other domains, and in useful ways?

>It's typically solved by fragmentation - allowing communities and
>organizations to develop systems which meet their needs, not a grand
>unified vision. 

Modularization, yes!
Restriction of interpretation, yes!


>Solutions which work (freight containers) are often
>standardized after the fact, but often address only parts of problems.

Sure, XML is no panacee for all your data problems. 
I would say it's a pretty useful mechanism to wrap-up data.


>In a sense, XML is out there, and various communities and organizations are
>partaking of it as they see fit.  Some like well-formed markup and don't
>care for the rest, others find DTDs add a useful layer of description, and
>others find DTDs are infinitely broken because they lack '&'.

This is what's great about it. The more general a tool, the more benefit for
more people.
It's just a tool. People use it in different ways. Script systems are quite
different. The alphabet is used in different ways. The English language is
used in different ways. That's not a bug.


>Markup is about the only fundamental thing I can find in all of this, and I
>think most people are reasonably content with that.

Markup is just an symbol encoding level.


>>Can we simplify?

>Sure, if you want to let people solve their own problems rather than
>forcing them into a single vision of problem-solving.

I for one wouldn't want to force people into a single 'vision'. Maybe that's
agnostic. I don't believe there is a single or best solution to any problem.

Sure, the alphabet restricts freedom but the trade-off works for me.


>>Can we re-express into solvable problem space.

>Locally addressable problem spaces, yes.  Globally solvable problem spaces?
>Maybe some of the locally addressed problems will prove to have similar
>solutions, but I wouldn't count on that producing unified answers.

Take this notion of local and global, and map that to my different encoding
levels. I claim that data can have several encoding levels. I claim that at
any level there is interpretation. 

Words do not have meaning, people give meaning to words.

Can we agree on that?
Words, tags, characters. These are all just different symbol packages.
Different interpretation levels.

If we want to communicate we have to restrict interpretation (and thus
define meaning). 

Apparently in the real world we are able to do this reasonably well. Good
enough at least to get some things done consistently. The more we restrict
interpretation, the more precise we can communicate.

Now my question to you was/is: at what level do you see a *critical* problem
for the viability of general markup language? 
Apparently you did spot a problem, and this is very valuable feedback. 
But please identify the problem. What is the problem? Why is it critical?
Then if we need to solve it, we need to restrict interpretation at some
encoding level. 
(You may remember my queries about STEP/EXPRESS.)

There is always a price to pay, I'd like to know if it's worth paying. 

Jan

Freedom vs. Restriction. Complexity vs. Simplicity. Chaos versus Order.
Randomness versus Structure. Noise vs. Signal. Thermodynamics.