HAL 0.1 release

Biju Chacko botsie at myrealbox.com
Thu Oct 2 08:29:25 EEST 2003

On Wed, 01 Oct 2003 21:33:30 +0200, David Zeuthen wrote:

>  But, I don't agree this is data proper, see my earlier post [1]. The
>  thing is that I still retain one of the goals of HAL should be to
>  support multiple OS'es with a single .fdi file. 

I haven't been following the HAL discussion very closely beyonding scanning the
mails, so I may be way off.

In general, received wisdom tells me that having code and data so thoroughly
mixed is a bad idea. I'm not entirely sure that scripting is necessary here.
While the flexibility it gives is nice, is it really desirable? Can the
requirement not be fulfilled by a structured data format.

For example:

(translate into XML, I'm too lazy to type all the markup)

setting = foo
setting.os1 = bar
setting.os2 = baz
setting.os1.variant1 = bun

The point is not XML vs python, it's code vs data. The conditional tests are
implied in the example above. Structured data *without* the explicit control
structures is less likely to be misused by the imaginative third party.

-- b 

Biju 'botsie' Chacko                        biju_chacko at vsnl dot net
http://www.symonds.net/~botsie          Public Key available on request

More information about the xdg mailing list