Re: [xml-dev] How to be nimble, agile in the face of changingtechnologie

[cbullard@hiwaay.net]

Variations depend on where you are in the stack AND the locale.  For  
example, if I started with that assumption, nothing I am writing would  
be of much use to the community in which I am situated which is  
squarely in the DTD camp and bloody big complex ones with multiple  
associated versions.  A mild bad surprise yesterday was to have  
converted a large RPSTL to a tag set only to open the DTD of note and  
find a different tag set:  same information, but different tags and  
different id'd elements.  Version creep.

Another variant is the stack:  are you writing XML software or are you  
using XML software to create an application (e.g., are you writing a  
parser or creating your own color-coded rich text box or are you  
dragging them into the forms, googling the API/docs and implementing a  
search and replace and a set of canned queries)?  Starting over has  
different costs.  In the latter case, throwing things at the wall for  
awhile until you are comfortable with the operations (eg, does it  
surprise you and do things you didn't plan for but realized you need  
only to find out it does that) isn't that bad.  Not checking the DTD  
first and completely understanding it? Not good but recoverable.  Not  
being told the customer switched to XML Schemas a week ago?  A bad day  
in hell.

Nimbleness is awareness and preparation.  Few get it right the first  
time and depending on the task at hand, no two developments start in  
the same place.  If you make assumptions based on "the larger XML  
community" then you are saying they are your customer or your  
supplier.  If you do that and they aren't one or the other, you're  
screwing up.

I'm building tools because I need them to do a specific job set.  YMMV  
and mine is a different job than writing for a market to which you  
haven't sold anything or which you propose to create.  On the other  
hand, if you are immersed in a particular market and what you create  
is saving you time and money, increasing quality and reducing misery,  
there is a better than even chance it will do the same for others in  
the same community.

Keep a copy of code that works even if you decide not to use it.  You  
never know.

len

Quoting John Cowan <cowan@mercury.ccil.org>:

> Costello, Roger L. scripsit:
>
>> In the case of the discussion at hand, the assumption is what? Is it
>> the following? Assumption: The larger XML community will always use
>> XML Schema so just create data models using XML Schema.
>
> Something like that.