Re: Subset Data Model

On 2012-08-14, at 9:24 AM, David Lee <David.Lee@marklogic.com> wrote:
> 
> This is my biggest objection to the development of uXML as well as the biggest objection to at least a dozen people to whom I have discussed it (but have largely been silent, I take it on myself to be their voice ... for now).

That is usually a bad idea.  I am confident that you can speak for yourself. I presume that may speak for MarkLogic, although probably not exclusively. I don't know that you have a mandate to speak for anyone else.

> 
> Without a clear definition of "What's the point" ... i.e. target audience, use cases, problems being solved etc., there is no objective basis for design decisions.  It is literally the cart before the horse.   How can you analyze the value of an implementation choice without having a clear definition of what it is you are trying to accomplish ?

David, the name "Micro XML" evokes, for me, an attempt to reduce XML.  That is, to eliminate the parts of XML that are no longer needed (think DTDs and entities), to incorporate those things that have become normalized (think XML:base, et al), and to generally create a profile of XML that is easier for more people. 

To the extent that we need a way to measure, I agree that this effort should enumerate its goals.  I am not convinced that the recent exchanges have helped move the group toward such a result. 

However, I will admit that I am operating on faith as much as science. That is, I have faith in the chair and editors to develop a document which I am going to be able to support.  

I think that your point is that the editors have not enumerated their assumptions. It is seems obvious to me that you are unlikely to agree with some of the editors' assumptions, but I do think that it could prove useful to enumerate some or all of them anyway.  



> 
> I suggest it is blatantly clear from the discussions that while many people think it is obvious "what is the point?" that no two people will give the same answer.  I have heard wildly different answers from the chairs and co-chairs of the committee itself ... let alone the brave members willing to ask.
> 
> Answering that question by ignoring it and skipping directly to development I suggest is a road to disaster.  You will not get consensus because people will not have the same idea as to what it is you are trying to build so their design criteria will differ.   Without consensus you might build something ... but what ?  That is a path of a single developer which is a fun and efficient way of building things ... often wonderful things ... but if you want community feedback, help, criticism and support and want to build something that has a measurable goal you must answer seemingly irrelevant or unanswerable and stupid questions like "What is it?" "Who is the intended audience?" "What problems are attempting to be solved ?" "How can we measure if the problems were in fact solved ?" ....
> 
> The alternative ... is to go build it, build the tools, write the documentation then come back and say "Here it is, take it if you want it" ....  That can (and has) led to wonderful things but it is orthogonal to a community process.
> 
> -----------------------------------------------------
> 
>

Received on Tuesday, 14 August 2012 13:59:57 UTC