Re: [xml-dev] costs of bureaucracy (was Re: [xml-dev] Not using mixedcon

["Simon St.Laurent" <simonstl@simonstl.com>]

On 4/8/13 3:30 PM, David Lee wrote:
> .. Taxes are the classic "There Is No Alternative" argument here.
> Are you really sure you want to use it?
>
> I am in no way referring to the tax *code* ... those are purely
> insane and have nothing to do with schema. But as Michael mentions, I
> am referring to , however insane the tax laws are, that I would much
> rather that the mechanism Of reporting be extremely constrained and
> not open to any old kind of data I, as an individual, wish to send
> in. This provides benefit to me as a user because it is more clear
> what is expected. This provides benefit to the bureaucracy as it
> reduces errors trying to fit the multi-shaped peg into what is
> inevitably *their* square box. And it reduces "creative
> interpretation" on both ends.

I don't think you've spent nearly enough time with tax forms or with 
businesses trying to decide how to make decisions and categorize those 
decisions to fit into specific boxes and their rather deep schemas.

Remember, I'm not calling for the abolition of tax forms.  I'm saying 
that you picked an example that I could have easily chosen had I wanted 
to make explicit the damage that schemas do.

> Seriously, are you proposing that every person should send in a tax
> report on a form of their choosing ? Or to be able to send
> electronically a filing in any data format they want ? How much money
> are you willing to donate to the human staff required to individually
> interpret all these ?   And do you trust them more then computers ?
> Like it or not, the whole *point* of bureaucracy is to produce a
> system that functions identically reguardless of whom is running it.
> In its day that was a very novel concept and provides a great deal of
> value to today's society, along with of course the inefficiencies of
> it. If you are seriously proposing that, then I will stop here as we
> live in non-intersecting universes. If you are *not* proposing that
> then what *are* you talking about ?

I'm saying that using the existence and enforcement of tax forms as some 
kind of proof that schemas are harmless is a really really bad idea. 
You brought it up.

Bureaucracies can impose forms on us because they have the power to do 
so.  There are memorable if occasional penalties for not complying.  Do 
you really want to put as many conversations as possible under the same 
constraints the IRS (for example) applies to its 'customers'?

I quite certainly support tax collection and general notions of fairness 
in such things.  I can also see that:

We depend on those mammoth bureaucracies to make things work, but 
despite frequent bursts of frustration about their many costs, the 
situation generally improves only marginally, and the number of data 
structures grows rapidly.

> Now on the other side of the fence, I work for a company which sells
> a "XML Database" (quoted because that's slightly not technically
> accurate, but close).  It is "schema free" in the sense that schemas
> are not required, and even if present don't trigger automatic
> validation. It is extremely useful to put "whatever you want" into
> the DB without any schemas.  Really nice.   But you can add schemas.
> Its also useful to either early on or later add schemas and make use
> of them where you wish. Not only does it support the type
> safety/coercions of XSLT and XQuery but some interesting things
> magically appear. For example if you just happen to have a schema for
> a document, even an incomplete one, then the XML to JSON
> transformations make use of that and apply the right type to say
> <weight>3</weight> and convert it to  "weight":3  instead of
> "weight": "3" If you decide later that "weight" should really be a
> string then you can make that change in one place (the schema) and
> the system adopts. I happen to like that.   And calling that a
> byproduct of a sick community is not really a useful statement.

Perhaps that statement is not useful to you, now.  I'm reasonably 
certain that over the long run, however, this conversation will be 
useful to the overall health of the markup community.

Thanks,
-- 
Simon St.Laurent
http://simonstl.com/