[QUANTITY] Quantity "arguments"

[Steve Lowe]

I think what is missing is requirements.

What I don't see in these discussions is a process that starts by 
listing in detail what the model has to do. What seems to happen is that 
we start with a vague idea of some desirable features, then jump 
immediately to a specification of classes and attributes. This starts to 
lock in design decisions too early, which leads to arguments. In many 
cases, the argument seems to be just about the *place* in the model 
where a given capability is locked in.

What is needed are written, detailed requirements derived from examining 
data analysis algorithms and existing data sets (use cases). The 
documents of this sort that I have seen so far seem superficial and 
short. Much more detail is needed. No doubt much of this information is 
out there in the heads of the folks in the group. But it is not out 
there on the table so that all the case can be in everyone's heads.

And if producing this documentation seems like a big job---well, yes, it 
will be. But why can't it be considered a tangible work product in 
itself, at least as valuable as the next demo?

Steve

Patrick Dowler wrote:
> While I'm sure all this is very interesting, it seems way off topic to me. I 
> am sure that's why only two people are left standing.
> 
> In the simple Quantity model that I put forward (somewhat privately, but that
> didn't last long :-( I tried to draw from all the available approaches and 
> ideas to make a "consensus". Of course, some ideas did not get included 
> because they were logically inconsistent with others that were more commonly 
> held, more valuable, more aesthetic, etc. 
> 
> The problem with this whole discussion is that it has degraded into "here's MY 
> model - see how I did it right" with a response like "yeah, but here's MYYYY 
> model, and it's pretty cool too". This gets us nowhere. 
> 
> At some point, people have to leave their models at home and provide 
> constructive criticism of the idea on the table. If something is missing, say 
> exactly what is missing. If something is wrong, by all means point it out.
> But bringing a whole new model to the table doesn't help.
> 
> So far, I have seen nothing on this list since ADASS that hasn't been said 
> before - either during the summer, at the Cambridge meeting, on a poster as 
> ADASS, or private communication. Yet my dm at ivoa.net mailbox is filled with 
> messages (on average 20KB!!!) not discussing the same thing.
> 
> my 2c (and 2c canadian isn't very much),
> 



-- 
Steve Lowe
Smithsonian Astrophysical Observatory
Harvard-Smithsonian Center for Astrophysics
slowe at cfa.harvard.edu
1-617-496-1661