Hi Adrian,
As I can understand the support of squid2.6 (important bugs
correction, help at squid-users mailing list etc) will not stop the next
day the squid3 will be released.
But adding new features to squid2 is something different. At this time
1-2 of the most important squid developers spend development time on
both squid2 and squid3 and this is huge overhead. This overhead must
eliminated. So the development of one of the squid2 or squid3 must stop
after a period. Squid3 looks OK, I do not believe that it will take 18
months to considered as stable from the users.
Which are the important squid2.6 features missing from squid3 and will
prevent users from upgrading to the squid3 ? Maybe these features must
be on top of the squid3 roadmap list....
Regards,
Christos
Adrian Chadd wrote:
> There's still an enormous squid-2 userbase out there which still needs
> to be looked after. Remember the fiasco over squid-2.5 + (lots of patches
> here) which people were running just to get "modernish" features?
>
> Squid-2 should be looked at as an oppertunity to incrementally test
> features out on a very larger userbase. Squid-3 and/or its children
> should be where these features end up, at least until Squid-3 matches
> Squid-2 enough for people to be comfortable migrating over.
>
> I think you're looking at a minimum 12-18 months after Squid-3 is seen as
> stable enough before people migrate from Squid-2 to Squid-3. I don't see
> this as any reason to cease Squid-2 development as it was practically done
> after Squid-2.5 was released - we should be actively bringing in features
> into Squid-2 that either have a Squid-3 counterpart (eg the IPv6 patchset
> Husni has against squid-2.6) or will be eventually ported over to Squid-3.
>
> Personally, I still plan on doing a lot of changes to Squid-2 to test out
> some ideas whilst finding some time to discuss, plan and code up Squid-3
> work. Squid-2 isn't going to go anywhere but if you just abandon it and
> force people to Squid-3 or bust right now they -will not- migrate and
> we'll just lose people to other proxy platforms. God knows we don't need
> that right now.
>
>> Looks like we have at least two Squid3 roadmaps already:
>> http://wiki.squid-cache.org/RoadMap/Squid3
>> http://wiki.squid-cache.org/Squid-3.1
>
> Well, I created RoadMap/Squid3 from /Squid-3.0; I think rolling /Squid-3.1
> into the Squid3 roadmap is a good idea.
>
>> I am afraid both Squid3 roadmaps are more fantasy- than reality-based.
>
> :)
>
>> I have outlined my specific suggestions on how to reorganize things in
>> the "v3.1 TODO and deadline" message on squid-dev. I am happy to edit
>> the wiki page (which one??) to implement those suggestions, importing
>> features on the current Squid3 roadmaps.
>
> It sounds like a good idea. I'm all for what you proposed in your previous
> email. I'm just hoping that we'll all agree on 3.0 -> 3.1 being a "clean
> up the insides and add IPv6" release rather than adding any more sweeping
> internal changes.
>
>
>
>
>
> Adrian
>
>
Received on Tue Oct 02 2007 - 13:13:16 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Oct 30 2007 - 13:00:03 MDT