XML Considered Harmful

[Dennis Lee Bieber]

On Tue, 21 Sep 2021 13:12:10 -0500, "Michael F. Stemper"
<michael.stemper at gmail.com> declaimed the following:

>On the prolog thread, somebody posted a link to:
><https://dirtsimple.org/2004/12/python-is-not-java.html>
>
>One thing that it tangentially says is "XML is not the answer."
>
>I read this page right when I was about to write an XML parser
>to get data into the code for a research project I'm working on.
>It seems to me that XML is the right approach for this sort of
>thing, especially since the data is hierarchical in nature.
>
>Does the advice on that page mean that I should find some other
>way to get data into my programs, or does it refer to some kind
>of misuse/abuse of XML for something that it wasn't designed
>for?

	There are some that try to use XML as a /live/ data /storage/ format
(such as http://www.drivehq.com/web/brana/pandora.htm which has to parse
XML files for all configuration data and filter definitions on start-up,
and update those files on any changes).

	If you control both the data generation and the data consumption,
finding some format with less overhead than XML is probably to be
recommended. XML is more a self-documented (in theory) means of packaging
data for transport between widely disparate applications, which are likely
written by different teams, if not different companies, who only interface
via the definition of the data as seen by XML.

>
>If XML is not the way to package data, what is the recommended
>approach?

	Again, if you control both generation and consumption... I'd probably
use an RDBM. SQLite tends to be packaged with Python [Windows] or, at the
least, the DB-API adapter [Linux tends to expect SQLite as a standard
installed item]. SQLite is a "file server" model (as is the JET engine used
by M$ Access) -- each application (instance) is directly accessing the
database file; there is no server process mediating access.

	Hierarchical (since you mention that in later posts) would be
represented by relations (terminology from relational theory -- a "table"
to most) linked by foreign keys.


-- 
	Wulfraed                 Dennis Lee Bieber         AF6VN
	wlfraed at ix.netcom.com    http://wlfraed.microdiversity.freeddns.org/