Re: [Aimmath-developers] my personal branch in the repository?

[Greg G. <gr...@ol...>]

Gustav W Delius wrote:
> 
> I have a computer at work and a computer at home and would like to do some
> AIM development work on both. CVS would be the perfect solution for this if
> I did a commit every time before switching computers. If I created a branch
> for myself and committed to that branch, then this would not interfere with
> the stability of the main trunk. Can those of you who are more familiar with
> CVS than me please tell me if they think that this is a bad idea or if there
> is an alternative solution? Otherwise I am planning to create my branch next
> Tuesday.

That's a good idea. With GAP we also have a development branch and a current
version branch. Most commits occur in the development branch, though
occasionally people will create another branch so as not to disturb things
elsewhere. The current version branch is only modified when bugs are
discovered and fixed.

I think we should have a current version branch and a development branch.
It makes sense for you to also have your own branch if you have stuff that
is not fully fleshed out and committing it would leave things in a broken
state.

What happens with GAP is that there is an attempt to `regularly' ship out
a new version. At the moment `regular' means `try for once a year'. In order
to keep to some sort of schedule, we now have an `essential' list and a
`wish' list, and a `freeze' date. A little before the freeze date we try to
determine the status of items on each list, and move anything off the wish
list for the next version that don't have any chance of fulfilment, and
determine how long it will take to deal with items on the essential list
to see if the freeze date needs to be moved. Anyway, we should put some
sort of plan like this together and decide when we want to release AIM 2.2
and what it should have in it. I have a few ideas for what I want in it.
I would like to see in it.

 1) Browser detection
    . if Mozilla/Netscape 6+ detected tth should use -u switch
    . if Internet Explorer detected no change, except that if IE 5
      is detected it should warn the user (student) that it is buggy
      and that multiple levels of indices are rendered wrongly
      e^(x^2) is rendered as e^(x*2)
    . other browsers warn that symbol font may not be rendered correctly,
      or else if we know they are unicode-capable use -u switch
    I have something modelled on the idea Manolis suggested a few weeks
    ago when I said there was a snag with using unicode with IE.
 2) Trig.mpl+
      This is a package I developed for some trig. questions. It's possibly
      a little too specialised and so I would like the other developers
      to look at it before it is included. Possibly some things can be
      pulled out and generalised.
 3) MultiRandom.mpl+
      This is a package I have developed to give feedback and randomise
      the order of the choices for MR and MC questions. It could probably
      be improved a bit further.
 4) Prob.mpl
      A small package I recently put together for a probability quiz,
      to automate checks for probabilities being between 0 and 1.
      In a week's time I'll need to make ready a normal distribution quiz,
      and no doubt will find some inadequacies of what's there currently.

... and there are a number of other minor additions and improvements.
Anyway, I'm not quite ready to commit this stuff. I need things to slacken
off a little so that I can tidy it up first.
      
  Regards,
  Greg