Re: [xml-dev] XML in the alleged Real World (was Re: [xml-dev] Does XML-Dev m...

[Eric Bohlman <ebohlman@earthlink.net>]

3/24/02 4:38:09 PM, "Thomas B. Passin" <tpassin@comcast.net> wrote:
>To judge by a lot of mail on the Mulberry xslt list, I would say there are a
>lot of people who are trying to work with xml and xslt without reading
>anything much about it, including the specs.  So they don't realize there
>are forbidden characters, and they don't know that xslt creates trees which
>may then be serialized, thinking rather that it is some kind of text
>formatter.  They don't know that &nbsp; is defined for html but not for xml,
>and so on.

What you're describing is the case where people are trying to use a tool without having a 
conceptual model of what the tool does.  That becomes a real problem when such people don't *want* 
to develop a conceptual model, but just want to know what magical incantation to type or path to 
click in order to achieve a result.  It's like trying to write programs without understanding 
concepts like variables, looping, or procedure calls.  It's very often accompanied by a macho 
refusal to read documentation, and a tendency to post questions to newsgroups or mailing lists and 
then get very petulant if the responses don't provide simple answers: "I want an answer, not a 
discussion.  Just tell me what to type."

For some reason this phenomenon is far worse with Web developers than other types of IT developers.  
It may be that these are people who got into the field because they really wanted to do graphic art 
and are frustrated because much of their actual work involves things they aren't really interested 
in.  Some of them seem to have an OhSoPoMo philosophical resentment of the idea that the syntax of 
something can be either correct or incorrect; in extreme cases, if you tell one of them "that 
syntax is incorrect" they'll come back with "in *your* opinion."  It's as if they're applying a 
completely esthetic paradigm to technology.

Developing a conceptual understanding or mental model of a technology does not, however, 
necessarily require reading the actual specs, which are generally written at the level of detail 
needed by an *implementor* of a processor rather than a *user* of a processor (and note that 
"developer" and "user" are not mutually exclusive categories; one stands in the "developer" 
relation to the application one is developing and the "user" relation to the development tools one 
is using to develop; they're roles, not identities).  But it does require more than just looking at 
examples and API references.