Abdel wrote:
To summarize my point: you will have to spend time studying the old
code before any comment becomes useful.
Not sure how to take that, but I agree if you only comment/document the
_differences_ as e.g. commit messages, instead of documenting the new
(intended) design. Then my belief is that the new design _should_ be
documented anyway.
So, unless you know already the old code, there's no point in studying
it; unless you want to waste time of course. Studying the new code
would be on the contrary very beneficial to me. That's why I said that
focusing on the cleanliness of the big patch is wrong. I stand on my
point.
On Tue, 14 Aug 2007, Asger Ottar Alstrup wrote:
The effort of producing an understandable patch series that will ease
the review costs so much that other developers with no experience with
that code can understand it, is just not worth it. The effort is much
better spent testing the result thoroughly, report and fix bugs, and
maybe add a bunch of comments in the header files.
I now understand Abdel's reasoning as follows[1]:
* The old code is very difficult and time consuming to understand
* Only you of the active developers understand it? [2]
* It'd be a waste of time trying to understand the difference as that
implies understanding the old code
* You therefore want people to be efficient by only looking at the new
code.
I wish it had been clearer earlier[1] so that we instead could have
discussed this approach in a better manner. I don't know if this approach
would be workable/acceptable, but let's at least start by understanding
Abdel's point of view.
Maybe we could forget these earlier discussions and start over?
What if Abdel post a new request for comments/discussion/review, asking to
check out your branch and review _specific_parts_of_it. I also think it
should be accompanied with an overview of the new design, with
references to the different areas in the source. [3]
Perhaps we are even starting at the wrong place. Rather than discuss the
new code, what if we discuss the new structure/design and it's motivation?
(This may well have been done before, but it might improve acceptance to
repeat some of it, rather than searching through the archives).
Please don't object with how this would be "inefficient"... That'd be
silly considering the time wasted so far in this thread.
Best regards
/Christian
[1] FYI, I didn't understand this from all Abdel's earlier posts. He may
been trying to say it, but it didn't come through to me at least.
I can of course only hope that I now understand...
[2] If someone else understands it, it'd obviously be good to ask those
persons look at the difference.
[3] I'm sure Abdel's written a good deal about this already, but please
don't say "look in my old posts". Abdel is the expert in this area.
He's already written the texts, so he's the best (only) person to
(re)summarize the new design.
--
Christian Ridderström, +46-8-768 39 44 http://www.md.kth.se/~chr
