3.6 Full release testing (was Re: [BUG]? Upgrade to full image

Martin Wirblat sql.mawi at t-link.de
Wed Aug 20 20:42:54 UTC 2003


Colin Putney <cputney at wiresong.ca> wrote on 20.08.2003 19:44:59:
>
>On Wednesday, August 20, 2003, at 05:13  AM, Martin Wirblat wrote:
>
>> Hi Göran,
>>
>> And if it is not in the image, it is not being used and so it will 
>> not get tested. That would be a big change in practice.
>
>Whatever gave you that idea? There are several packages on SqueakMap 
>that I use *constantly.* I'm sure I'm not the only one. You seem to 
>think that if something is not in the image or the update stream it 
>doesn't exist. This isn't true.

But you are using them in your image, right? I just wanted to show 
that Göran's logic ( that you snipped ) may be turned around. 

>> The Full-stream would give us a cheap testing-setup by giving as 
>> many people as possible the chance to play around with Full  before 
>>  the release, even if we can't test everything 100%.
>
>I don't see how putting something in the update stream has any impact 
>on how well it's tested. People will test things that are important 
>to them. If they are important they'll get them from SqueakMap just 
>as readily as from the update stream.

1. There is the effect of what I called playing around. Or to put it 
differently: We all are controlled by something called subconscious, 
believe or not. If for example something is achievable only with work, 
this subconscious may not even let the idea of doing it coming through 
to the consciousness. So something which is done easy, will be done 
more often. If there is no easy update, more people will wait for the 
final release. Let's think of a community testing statistically while 
having fun. 

2. Even if you don't play around, because you *use* packages 
*constantly*, you will -as you said- use them constantly, meaning you 
will not upgrade every now and then and thus you will not be testing 
them. Of course to use them ( in a stable manner ) is perfectly valid, 
you are just not verifying their evolution. So this is a different 
story. But if you _want_ to be a tester for those changing packages in 
a changing environment, it is convenient for you to have an 
incremental, automated update tool. 


>If they're not important, they won't get tested. And that's a good 
>thing. Let's not waste time testing stuff that nobody uses.
>
>Colin

If we release something it must be tested almost by definition, 
otherwise it is no release, it is more of a joke. If something is in 
Full which nobody uses, it should be removed from Full. 

But 'releasing' Full and testing only Basic somewhat means effectively 
getting rid of Full. I have the impression that some people's 
subconscious is driving their opinions to this thread in this 
direction :-) 

I guess this is not what the majority of Squeak users wants.

Regards
Martin 



More information about the Squeak-dev mailing list