On Feb 11, 2009, at 7:34 PM, Keith Hodges wrote:
Stéphane Ducasse wrote:
Where? When? Keith in the long run it would be nice that yu do not bash us systematically. Yes we have a different way of working but normally this is often by pure lack of time or information that we do not check what exist.
Stef
I was concerned to find code in Pier that did not seem to be compatible with squeak, while I thought that the Workspace issue had come up before.
no idea.
Your different way of working..... as I see it:
You want more tests, you develop more tests, you add more tests to the image. Are those tests useful to the rest of the squeak community? No. Should they be? YES.
If there was a way of marking those tests that are pharo only tests, as such, and also marking those tests that should work in squeak. Then the squeak/pharo communities could use one testing package for all squeak images, and pharo.
may be
So to enable this to happen you need to think about testing in a broader sense and you need improved tools to handle it, and you need to develop or assess those tools FIRST.
Yes.
hmm, so August 2006, I thought this through, that to support diversity in the squeak community we needed something like that, and SUnit-improved was the result. That was its purpose, to support forks of squeak, I believe in forking squeak....
I know I should have a look at SUnit-Improved but my time is limited
All the tools I have worked on for 2 years are to enable working with diverse images, and to facilitate packages to work in different images. Thus reducing the dependency upon a specific kernel-image and thus releasing the kernel-image to be improved.
What I don't believe in was the pharo approach of going your own way WITHOUT considering anyone else, to the extent that you ignore all of the extensive work that has been carried out to provide tools that would enable you to participate. The main goal of the 3.11 effort hasn't been a particular image, but a tool set, and a process, that would enable folks to work on diverse projects, while considering each other.
Our goal is to make progress. Squeak is dying don't see it?
And you are ignoring those tools, because you dont see them as important.
No. The problem is the noise ratio with them (nul pattern, bad formatting, extensive extensions....)
I cant work out why you dont see them as important is because you have sucessfully harnessed a lot of enthusiasm for improving squeak. YET you dont have a process for using that enthusiam effectively. You still have the same old bottlenecked controlled process where there is no actual plan (you do what you feel like when you get up in the morning), and two people get to hack at the image day and night and everyone else is contributing to a moving target.
I would love to have a buildServer and also run automatically SmallLint rules on the code now either pharo does not exist because we build those tools or pharo exist. So we are going slowly on both side. I was hoping to get such buildserver and this is why we started to write on Installer (and you know that this is not only for pharo so you should be happy). Now you should ask yourself the question why people are not all using your tool extensions. Conservatism? stupidity? lack of time? lack of trust? Dont'care? Stef