On Wed, Jan 21, 2004 at 01:59:21PM -0500, Meng Weng Wong wrote:
| On Wed, Jan 21, 2004 at 01:53:07PM -0500, George Schlossnagle wrote:
| | Besides consuming about 3 times as much space, what does this really
| | buy? IMHO 'human-readable' data formats are over-rated when their
| | consumers are all automated processes.
|
| In theory, extensibility will be built in.
|
| To extend SPF you need to put an RFC through and spec out v=spf2, which
| takes longer.
|
| I am not familiar with XML so could an advocate please give a shout.
|
| If we weren't dealing with 20 year old technologies we would agree in an
| instant that XML was appropriate.
XML is a markup language. It is intended for documents to supplement them
with meta data, such as aspect information, formatting suggestions, and such.
Much of what XML is being used for, such as outputting database tables, is
not that of a markup language. It's an abomination.
The reason that XML got adopted for things like that is because there was a
lack of any well known standards. One did exist back in the 1970's but was
largely ignored due to lack of networking to do anything with it, and now
it could be used but almost no one knows of it (too few were even born then).
It was called something like HDF and looked very much like the format used
for BIND version 8 and 9 named.conf files. XML just happened to be there
with already coded (although rather obese and buggy) tools to break it up
into structures. It's those Java programmers that are at fault.
XML has certain philosophical inconsistencies. It claims to be a human
readable language, but that is only to the extent it uses printable text
characters. The text itself is very hard to read, and harder still to edit.
XML has validators, which as far as I can tell are useful only in cases
where humans tried to edit XML files (it's so easy to goof them up), or
when humans coded XML generators incorrectly (everything would break if
that happened anyway).
I run http://linuxhomepage.com/ which inserts news feeds from many sites.
About half come in an XML flavor like RSS or RDF. XML parsing tools are
really a major pain to use because there are constant problems with the
data failing the validation. 95% of feed failures are due to problems with
XML and/or its implementations.
| Think of the people you know who've implemented SPF. Would they have
| done it if they'd had to do it in XML? Assume the XML approach had all
| the auxiliary support: a wizard, education, that sort of thing too.
I'm working on the design of a very lean LMTP/SMTP daemon to receive (only)
mail coming in. It will need to have SPF capability among other things.
Requiring XML parsing will basically ruin that leanness.
Please keep SPF "lightweight" and make it something that can be parsed in
very few lines. Once parsing tools are written and exist for various major
languages (C, C++, C#, Java, Perl, Pike, Python), why do we need XML? It
would not be to get read-written parsing tools. And SPF is already human
readable, so it would not be to get that, either.
Maybe XML is being proposed because it is still the "format du jour".
If anything, SPF's data model is more like that of the Forth language.
--
-----------------------------------------------------------------------------
| Phil Howard KA9WGN | http://linuxhomepage.com/ http://ham.org/ |
| (first name) at ipal.net | http://phil.ipal.org/ http://ka9wgn.ham.org/ |
-----------------------------------------------------------------------------
-------
Sender Permitted From: http://spf.pobox.com/
Archives at http://archives.listbox.com/spf-discuss/current/
Latest draft at http://spf.pobox.com/draft-mengwong-spf-02.9.4.txt
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to
http://v2.listbox.com/member/?listname(_at_)©#«Mo\¯HÝÜîU;±¤Ö¤Íµø?¡