Re: [xml-dev] Create fake stuff (that’s all you can really do anyway)

[Thomas Passin <list1@tompassin.net>]

This is very wrong in a lot of ways.  It's also right that one is 
usually thinking in terms of some model, and that model may be closer or 
farther from the reality of nature.

What's off target is the role of an XML schema.  What is it you have to 
do here?  Develop a weather model, or construct a schema or interchange 
of weather information?  They are not the same thing.  Developing a 
weather model requires a lot of physics.  Developing a model of the 
information needed to be exchanged requires knowledge about modeling 
structured information, and a much smaller amount about physics.

I spent several years, on and off, in a similar position working on 
schemas for (vehicle) traffic information, and I didn't know much about 
traffic.  But we started from preliminary schemas that were already in 
existence, and worked to correct and improve them (actually, we also had 
to do ASN.1 <==> XML conversions).  Yes, we had to try to do some 
modeling of traffic so that we could understand how the XML messages 
supported by the schema would work.  And we had some idea of what kinds 
of information people needed to exchange.

In the case of weather, or weather patterns, the request for a schema 
wouldn't happen in isolation, divorced from any needs of scientists or 
meteorologists.  If it were, it would be like asking for a specification 
for a new airplane, but without giving any indication of what the users' 
requirements were.  If that were the case, yes, give up!

Otherwise, take what you get from the users about what they need, and 
carry on.

TomP

On 1/21/2017 3:13 PM, Costello, Roger L. wrote:
> Hi Folks,
>
> So, you’ve been tasked to create a model of weather patterns. Your
> deliverable is an XML Schema. People will then use your XML Schema to
> generate XML documents and exchange those XML documents.
>
> Ah, the world will be so much better with the ability to exchange
> standardized weather data.
>
> But how will you model weather? What will your XML Schema contain? You
> do not have expertise in weather science. How long will it take to gain
> the expertise to create a useful weather model? How deep into the field
> should you go?
>
> Answer: Give up. You have no hope of ever understanding weather. Even if
> you are already an “expert” in weather science, you have no hope of
> truly understanding weather.
>
> Allow me to explain.
>
> But first, a detour.
>
> …
>
> 1 + 1 = 2
>
> Obvious, right?
>
> Not at all. It took two of the most brilliant 20^th century
> mathematicians (Whitehead and Russell) 300 pages to prove that 1 + 1 = 2.
>
> I have this note next to my desk, which I read every day:
>
> /Everything is vague to a degree you do not realize
>                 till you have tried to make it precise./
>
> Scientists, mathematicians, engineers, philosophers, sociologists, and
> others recognize the futility of tackling reality head-on. So instead
> they invent miniature – fake – worlds, containing small sets of fake
> symbols and fake rules. They then manipulate the fake symbols, according
> to the fake rules, to see what kind of fake results they get. Sometimes
> there is a correspondence between the symbols and rules in their fake
> world to things in the real world, and so the results they derive in the
> fake world are then applied to the real world (with the caveat that it
> might be completely wrong).
>
> Don’t be deceived (or have the conceit) to think that you fully
> understand something. You don’t. You’re understanding is appallingly
> shallow. Mine as well, of course.
>
> …
>
> Okay, back to the weather XML Schema. You will never understand weather.
> Abandon such ambitions. So what to do? Answer, invent a miniature – fake
> – world, containing a small set of fake symbols and fake rules. Create
> an XML Schema for that fake world. If you’re lucky, exchanges of your
> fake stuff will have utility to others.
>
> Comments?
>
> /Roger