Posted by TimP on Sat Oct 31 14:09:35 1998
In Reply to: Re: Quora (group bonding) posted by william on October 30, 1998 at 20:51:12:
| Distribution: | paneris@i-way.co.uk |
|---|
>> The essense of it is that 'owned code' is bad code.
>> Just as the compiler is your friend so is the reviewer.>The difficulty here is that code has to be very good before
>it is worth other people using it.Not so, it just has to be good enough:)
>This means that the only unproblematic way that code sharing
>and code reviews can happen is over the long term, with
>a body of material which is updated over several projects.>Now in fact between Paneris, the PM system and Iglu we
>have a pretty substantial body of code and also now the
>leisure to review and share it.> But I predict that unless
>we work at it it will not happen.>Perhaps we can play
>a game in which we have to rip our code to shreds, and
>everyone else has (honestly) to try to defend it :).I believe that the best way is to hand over ownership of code.
This actually removes the slightly artificial 'review' step.
I have battled with other peoples code, and as I do I change variable names
to reflect my understanding, add comments and try to
answer any queries left in the code.I have been trying to get shared authorship to work on the Panris site,
and have worked on code originated by everyone except Myles,
though he left a trace of his passing on the muliboard code.To get shared authorship to work people have to have the ability and
courage to take over where someone else has left off.To have the ability to change someones code we have to ensure that
we all have the same tools and setup. We also need to have local test systems
to work upon.I wish to see everyone with a local copy of paneris (the site and scripts).
My main stumbling block remains the lack of a working PMS system at 77 Hurst street,
which would enable me to review and revise the PMS system.This is, from my perspective, even more important than a change control system.
yours
timP