[getopt-sig] The bake-off

[A.T. Hofkamp]

On Thu, 30 May 2002, Moore, Paul wrote:

> Sorry, you're right. I think I was thinking of Albert Hofkamp, who came up
> with an alternative to Optik. But as I said, I never quite followed what he
> was trying to suggest was the benefit for his package compared to Optik.

I wasn't :-)

I was trying to understand the core issues that need to be dealt with, when
writing an option parsing library.
I have seen a quite large number of (C++/Java/Python) option parsing libraries,
and all of them just seem to grab a solution from the air, implement it,
and present it as the solution (Optik does that too).

Yet for some reason, getopt is the most commonly used option parsing library.

Therefore, I concluded, there is something that getopt has, what is missing
in all the other libraries.
I thought (and still think), that in order to write a good option
processing package, we need to dig what that something is.
Without that knowledge, the only possible result is an option processing
package that also nobody uses.

In other words, I haven't yet arrived at the point where I understand what
an option processing package needs, let alone have implemented something
generally usable.

argtools was/is just a case study for this SIG (although the tool is actively
being used here), and in a sense some counter-weight for the dominant
presence of pro-Optik.

> a sign that Optik had "won" than anything else. I would have expected the

I considered it a dead-lock, as explained in my previous post.


Albert
-- 
Unlike popular belief, the .doc format is not an open publically available format.