> > On Thu, Mar 04, 1999 at 12:04:21AM +0100, Asger Alstrup Nielsen wrote:
> > > After Joacim's angry and furious rant about the lack of design documents,
> > > I figured that I would humour him a bit and write 11 pages about the
> > > strings and encodings in the new kernel.
>
> Asger I find this paragraph particularly funny after having just read your
> email in the libsigc++ thread asking them to write some design documents
> even if they have to be changed frequently.
Thank you ;-)
I hope everybody realizes that I really agree with Joacim, but just hate to
admit it.
To defend myself a bit: There is a fundamental difference between a library
and an application. In an application, it's really important to document all
the features the program offers. The Doc Team handles this for us.
When it comes to a library, the primary audience is other developers, and
therefor, the documentation of the features in the library is alpha-omega.
> You've got 'merged words' enabled or whatever the ispell setting is.
> That is, 'there' + 'for' = 'therefor' which must be spelled correctly
> because joined words are accepted.
As I wrote to Amir, the rationale begind writing "therefor" is Danish spelling.
The word in Danish is "derfor". Much shorter, much easier to spell, and much
better. Derfor, I'll just use the Danish word from now on.
Amir:
> I would call it a Pyrrhic victory (but only because I'm trying to show off :)
I just looked this up: "A Pyrrhic victory is a victory that has cost way too
big losses for the victor" (or something like that.)
Nah, it's not such a victory. I enjoyed writing it, and didn't loose anything
by doing it. All the stuff about experiments and so fort was just a marketing
stunt to get some attention. It seemed to work ;-)
> I'll get around to reading it eventually but since we've had several long
> discussions on the subject over the last year and a half I'm hoping Asger
> made use of cut and paste from those discussions to build his document --
> and hence it would only be revision for those of us who read them then.
Of course it is based on the discussions we have had over time. However, there
are a few things that are different from what we have discussed earlier, so
consider the design one step further in refinement.
There should be no really big surprises, but I think there are a few points
worth reading, also for the experienced developers. (The real and ugly truth
is that I attacked this area after having seen Lgb's comments in the code where
he asked for clarifications. And when I wanted to write those clarifications,
I realized that the design could be better. So I started a small document
about it, and suddenly it grew to those 11 pages.)
> That doesn't mean its not important. After all, newbies can read and
> such documents can get them up to speed about our design decisions
> (whether they be collective or individual decisions).
Agreed. And devvies can too. For instance, your fine documents about the GUI
independence gives me a freedom to be more surveying than participating in some
of those discussions. I trust you to sum up the things and keep the document
fairly up to date, and in this way, we can track each others efforts without
all doing them.
Yes, I'm saying that I do not really mind if people do not comment on the
design documents that much. I trust people to read them, and complain if they
disagree.
Greets,
Asger
