[Sigia-l] tenets, principles, standards, etc ...

[Listera]

"David Heller" wrote:

> The brackets are one version, w/o the brackets is another. Our customers
> said the brackets is the standard they deploy with.

David, that's slaughtering the word 'standard.' That's *the* way *this*
particular client is using it. In the grand scheme of things it's just *one*
way. There's nothing 'standard' about it.

> Why is this important? B/c if we worked w/ the bracketed standard it meant
> that we were going to create a rule to help users support this which would
> basically force unique category names across all the taxonomy instead of
> unique category names at a container in the taxonomy.

Without going into already-covered territory of RDMS here, there are other
ways of storing and displaying unique/non-unique nodes of a taxonomic tree.
 
> Is uniqueness across a taxonomy a standard? Is it a valuable one? At the
> time of the discussion there were mixed opinions on the subject. What
> was clear back then was that the value of the standard was not very well
> understood, or it's value was based on the assumption of very poor
> technology.

And one sure way to inflict pain on consumers for years to come is to pick
one of these, declare it a 'standard' and shove it down their throat.
 
> Another example of an IA "standard" would be JJG's visual vocabulary. It
> is a stencil set for use for making flowcharts with for both content
> based and application based situations. Very cool actually. It could be
> standardized in an organization so that there is ONE method for doing
> flow charting and so that the key for a flowchart can be taught once
> instead of constantly re-invented.

It's way too complicated for non-geeks. For non-techie consumers like
marketing, management, business-line people, etc., a much more simplified
version is needed. I actually devised one with only four branching symbols
that covers something like 85% of the territory. It's called SimpleFLOW. I
never released it, I still think it should get even simpler.
 
> Should it be a standard for ALL IAs? No way. Maybe a base or guideline,
> but that's about it.

Fair enough. Look at your project/goals, think through, adopt what's
available. Why the need for a 'standard'?
 
> Why are standards important? Or should I say when (at what level of
> organization); when interoperability is at stake and that
> interoperability derives explicit or implicit efficiencies and value for
> that level of organization.

Interop is hugely important. But there are critical differentiators here.
One camp (guess which:-) says you must use our OS/apps/glue/etc, they all
work together. The other camp accepts the world is heterogeneous and tries
to connect them at mutually agreed upon format-level commonalities.

An app server not hobbled by lock-in can interop with any client (web
browser, Flash, mainframe, PDA, etc) over HTTP, SOAP, XML, etc.

How did we get here? Mainly because IT bureaucrats locked themselves and
their users into internal 'standards' without thinking through long term
consequences of their actions. The inevitable aggregation of such seemingly
innocent 'internal' standards sooner than later stagnates the whole
industry.

Even you agree that Visio has some serious shortcomings to be the 'standard'
for IAs. But it's exceedingly difficult to compete against Visio to provide
better tools because so many standard pushers have declared it the tool you
must use to be a part of their team.
 
> This speaks best to Ziya's distinction between tools and formats. If
> there were proper open formats to achieve what is necessary here, then I
> could say, "Cool!" I agree. But there are no open formats that I have
> seen that are completely interoperable across the tools that use those
> formats.

Yes, but for the love the-fat-Santa-with-a-big-sack-of-presents, let's work
towards creating/promoting open format-level standards rather than declaring
our own internal habits/preferences 'standards' and trying to
license/bureaucratize a necessarily-hybrid field.

Ziya
Nullius in Verba