1 |
On Mon, 18 Jan 2010 08:59:07 +0200, Alan McKinnon wrote: |
2 |
|
3 |
> Most devices fall into one of two groups: storage and I/O. |
4 |
> Auto-mounters do not care about your keyboard, whereas X needs to know |
5 |
> about your monitor, card, keyboard, mouse. Why does hal try and |
6 |
> abstract both? Seems silly to me. |
7 |
|
8 |
On the other hand, having a single method of configuring such things does |
9 |
give consistency, and means you have to learn only one syntax, but see |
10 |
below. You cannot totally separate the two areas, for example a keylogger |
11 |
may need access to both I/O and storage, so a central, separate resource |
12 |
used by all software is more in keeping with the Unix way than each |
13 |
program including its own implementation. |
14 |
|
15 |
> One could also argue that the developer's state of mind is reflected in |
16 |
> the chosen method of configuration - xml files. This just defies all |
17 |
> understanding. Apart from the fact that real-world xml is almost |
18 |
> unreadable, the conditions that make xml useful are simply not present |
19 |
> in hal... |
20 |
|
21 |
I couldn't agree more. XML was very fashionable a few years ago, maybe |
22 |
this influenced the developer. Hell, I was even guilty of using it |
23 |
myself :( As an alternative to binary configuration files, XML is a step |
24 |
in the right direction, but it should not be used where users are |
25 |
expected to edit the files. In some ways, the worth or otherwise of HAL, |
26 |
from a user perspective, has been largely obscured by the difficulty in |
27 |
reading, let alone editing, its configuration files. |
28 |
|
29 |
|
30 |
-- |
31 |
Neil Bothwick |
32 |
|
33 |
I am MODERATOR of BORG. Follow the rules or be assimilated. |