[
Lists Home |
Date Index |
Thread Index
]
>Not for people who want rich text processing. As I read your
>response, it makes me think that the only parts of my message you
>read are the little snippets you quoted. I'm not sure how you
>managed that, but you might want o respond to the whole concept I
>presented and not just the sound bites or we'll be trading
>quiddities all day.
Ok, I'll re-read and re-spond. In then meantime...
>> >Comma delimited files provide *all*
>> >the power of XML for relational and OO dumps with much fewer
>> >inefficiencies and much less cruft (CDATA sections, entities, etc.)
>>
>> I disagree - the structure of csv might be a good match for
>relational data
>> *if* there was a convenient way of making links between files. Ok,
>> hyperlinking (XHTML, XLink or whatever) isn't strictly part of
>the XML spec,
>> but if you stand back from the spec you see a whole load of layered
>> technologies. If you stand back from comma delimited files all you see is
>> smaller comma delimited files.
>
>Are you seriously saying people wouldn't know how to layer
>structured linking on top of CSV? I find this impossible to
>credit. I have seen the trivial solutions to this many times in CSV.
That isn't what I said at all. My point was that there is already quite a
degree of sophistication in the XML specs, which I am not aware of existing
for CSVs.
Structured linking is but one example, one which comes in handy for dealing
with relations. I might have chosen something from the OO domain like
inheritance, or encapsulation though I'm sure you've seen the trivial
solutions to those as well. These things have been worked out to varying
extents in XML, but where are the "..the standards gold, vair, purple,
opposing..."?
>> The structure of XML is hierarchical (element tree) and has
>property slots
>> (attributes) and so makes a much better fit for OO data than flat files.
>>
>> >Again, we are here talking XML, not OO or relational systems. XML
>> >is a *text processing system*.
>>
>> Rather a pointless statement IMHO, like saying a relational system is a
>> binary processing system.
>
>It *should* be a pointless statement. As you suggest, it should
>be somewhat of a tautology. But if you don't know why I had to
>make that statement, anyway, you either haven't been around here
>very long, or you haven't been listening.
I don't really think it is a tautology though. Ok, a major aspect of XML is
its text representation, but most of the interesting stuff happens at other
levels. I've been around long enough to realise that there is more to a
document than the words.
>> My potted theory of why XML has
>> >gained so much sudden prominence in such a diversity of fields is
>> >that a lot of the weaknesses of the super-platonic approach to
>> >software design (i.e. relational and OO thinking) have been
>> >showing of late, and developers were initially amazed at how much
>> >more productive they could be with loosely-coupled systems based on XML.
>>
>> I'm not entirely sure which aspect of platonic thinking you have in mind
>> (!clintonic?) but I think I agree here - XML more than anything
>has acted as
>> a catalyst for interop, breaking people out of closed systems.
>
>Never mind the philosophy, then. I read Marsilio Ficino and his
>lot with a grim fascination in college. I've never been able to
>shake the experience. Replace "super-platonic" with
>"based-on-artificial-imposition-of-design-uopn-nature".
Ok, thank you, that's clearer.
Cheers,
Danny.
|
