VOEvent - Param comment

[Tony Linde]

<<All design is evolutionary, therefore all technologies are dead-end.>>
 
I meant dead-end in the workflow sense - once you produce a VOTable, it has
to go back to the user for interpretation - it is the end of the workflow
(unless the user tells the next step in the workflow what they think might
be in the VOTable, but not being based on schemas, this is a risky thing to
do). The same applies to the param parts of VOEvent.
 
<<cause the standard to evolve>>
 
And that is my concern. Instead of abandoning the param tag, people will
fudge/kludge it with something like the utype attribute, making it an even
worse construct.
 
<<the schema approach is the more pragmatic
That may be the first time this statement has been made :-)>>
 
Yes, I know I am endlessly repeating myself. I'll stop and let Petr carry on
:)
 
T.


  _____  

From: Rob Seaman [mailto:seaman at noao.edu] 
Sent: 07 August 2006 22:02
To: Tony Linde
Cc: voevent at ivoa.net
Subject: Re: VOEvent - Param comment


Hi Tony, 



The reason that VOEvent and VOTable has stuff like <param> tags and the like
(and VOTable has a lot worse) is because they are dead-end technologies.




Let's give it a few years and see :-)


All design is evolutionary, therefore all technologies are dead-end. No
natural selection without obsolescence.



They cannot be used for anything except being looked at by humans



Even this would be an improvement on the state of astronomical alerts before
April 2005. If, however, VOEvent packets do prove unusable for autonomous
purposes, I am confident that the rather aggressive interest in driving
these initiatives forward will cause the standard to evolve in response.



(which is the most daft fudge I've seen in over 30 years of development)



Sounds more like a kludge than a fudge. Cold fusion was a fudge.



the schema approach is the more pragmatic



That may be the first time this statement has been made :-)


I think we would all love to see the schema approach succeed - just like
we'd all love to see ontologies prosper like the fruits of Soviet five year
plans. Assertions about pragmatism are themselves rather devoid of semantic
content. Either these things will succeed or they won't. Whatever happens,
it's hard to believe that one size will fit all situations.



I will say that at least VOEvent does have its core data in properly tagged
elements



Given this comment, it sounds like we got the balance just about right!


The point was to render unto Caesar (if only Caligula, not Julius), what
needed to be rendered. We deemed the citation mechanism as key to follow-up,
for instance, and focused on limiting the semantics. We perceived space-time
coordinates, the independent variables of an observation, as requiring
special handling and pressed forward with STC. This was not done lightly :-)
And we also decided that the general question of how to handle the
completely open ended situation with dependent variables was best handled
using the <param> mechanism, rather than by requiring a complete a priori
data modeling exercise for every event type.


Will our duckling be revealed a swan - or merely fowl and homely?


Rob


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/voevent/attachments/20060808/3f0f5912/attachment-0001.html>