ducasse stephane ducasse@iam.unibe.ch wrote (in a pvt msg)...
Hi dan,
I was wondering if it would not be wise to have an official version of Squeak now. Like that we can have the time to work on the new versions with modules.
Else the module process from my experience may take some time and having a new version now could be referred as the version before the modules. Especially because 3.0 was put together in a hurry and is not really good for demoes and the like. For example I'm writing columns and I'm referring to 2.8 because 3.0 has broken examples and I would love to move to 3.1. In addition this also give a new release which is good (like microsoft) people have to upgrade. As an example I could put 3.1 on the cd with the article and say the new Squeak ;).
Hi, Steph -
Ahh... a message I agree with. (;-).
Yes, I think you're absolutely right. I'm wedged with a couple of other projects, but in the next day or two, I'll issue Doug's most recent harvesting (35 updates) and seal off 3.1 alpha. Requestors of updates will be given a choice of going to 3.2 alpha, or settling in to 3.1 beta. We'll keep the 3.1 beta updates open for a week or two, and then declare it 3.1 final.
Incidentally the main failure with 3.0, in my opinion, was that no one took full responsibility for the final phase where fixes and nothing else should have been carefully added, and then a final clean image posted. This is a place where SqF, and the harvester-types especially, could be a great help. It is my hope that a few people who are motivated to see a nice stable 3.1 release will organize themselves and help out by selecting and testing all updates appropriate to 3.1 after it goes beta. Given such help, I am willing, as time allows (meaning probably after OOPSLA), to do all the normal work of tidying up the image once we have settled on the final updates (I'll probably send out a gamma image for this purpose near the end).
Following the step to 3.2 alpha, I will issue the module updates, so 3.2 will be (at least "weakly") modular from the start.
- Dan
Hi dan
Hi dan,
I was wondering if it would not be wise to have an official version of Squeak now. Like that we can have the time to work on the new versions with modules.
Else the module process from my experience may take some time and having a new version now could be referred as the version before the modules. Especially because 3.0 was put together in a hurry and is not really good for demoes and the like. For example I'm writing columns and I'm referring to 2.8 because 3.0 has broken examples and I would love to move to 3.1. In addition this also give a new release which is good (like microsoft) people have to upgrade. As an example I could put 3.1 on the cd with the article and say the new Squeak ;).
Ahh... a message I agree with. (;-).
But agreeing or not has nothing with contents..... Ok it has ;) If you are refering to {} we should have a discussion with arguments once. I was discussing with nathanael and roel and we ended up with the same conclusion. So I would like to see the other camp point of view.
I think that Scheme has a process for improvement that allows parties to write down arguments and confrontation. May be they have too much time to be able to do that ;)
For the module decomposition it would be great to see if we can use the analysis of Ginsu, this way you could check a decomposition. Hans-martin is far better than me so I'm sure he will come up with something ;)
Yes, I think you're absolutely right. I'm wedged with a couple of other projects, but in the next day or two, I'll issue Doug's most recent harvesting (35 updates) and seal off 3.1 alpha. Requestors of updates will be given a choice of going to 3.2 alpha, or settling in to 3.1 beta. We'll keep the 3.1 beta updates open for a week or two, and then declare it 3.1 final.
Great!!! I think that this is important. we will be able to push Squeak a bit more with 3.1. It would be great if we could have Squeak 3.1 for mid november like that we can push it for christmas ;) I hope to have a nice column with 3.1 as gift. I have a friend who is finishing a french book on Squeak and using 3.1 instead 2.8 would be really important for him.
Incidentally the main failure with 3.0, in my opinion, was that no one took full responsibility for the final phase where fixes and nothing else should have been carefully added, and then a final clean image posted. This is a place where SqF, and the harvester-types especially, could be a great help. It is my hope that a few people who are motivated to see a nice stable 3.1 release will organize themselves and help out by selecting and testing all updates appropriate to 3.1 after it goes beta. Given such help, I am willing, as time allows (meaning probably after OOPSLA), to do all the normal work of tidying up the image once we have settled on the final updates (I'll probably send out a gamma image for this purpose near the end).
I want to test the image as soon as it is out. I plan to redo all my demoes so I can certainly check some parts.
Following the step to 3.2 alpha, I will issue the module updates, so 3.2 will be (at least "weakly") modular from the start.
I do not know if you want to play the same trick that with 3.0 with having two tracks because I imagine that introducing and refactoring code may shake the image for a moment. But I'm not sure that having two streams is good either.
Stef
squeak-dev@lists.squeakfoundation.org