[TYPO3-core] Proposal for the upcoming Roadmap and LTS

[Georg Ringer]

Hi,

Am 06.11.2012 22:18, schrieb Mario Rimann:
> Or how do you sell these upgrades to your customers? Do all of them have
> a payed SLA with you that cover your work to upgrade each and every of
> your customer installations from minor version to minor version
> throughout the year? I can't imagine that.

a customer should update if there are features needed and or there is no
support more for the version.

> When I read this mail and from what I've discussed with some of the core
> devs in the past months, I fear that at least some of you have a very
> different viewing angle onto this whole topic as some of the users
> (agencies and their customers) have. 

I guess most core devs work at agencies and are not locked somewhere
just doing strange stuff without ever seeing customers or the daylight. ;)

> I personally like the LTS version a
> lot, as it gave us and our customers a lot of stability - and to be
> honest, the LTS version is also a selling argument!

true but we would always start with the current stable version, e.g. now
4.7 if it is a bigger project as it got more features we need and there
will be also less pain to update to later versions

> - If the same user would buy his website now (Q3/2012), his Installation
> will have an expected lifetime of just about 1.5 years - almostly
> independed of whether he gets a 4.5.x LTS based TYPO3 setup or a 4.7.x
> based version (differend of the two EOL dates is 6 months)

yeah but isn't that always the case in daily life? If you buy now the VW
Golf v6 you will have a new car for you but there is already the version
7 out.

in earlier days you earned money by selling DAM, with 6.1 you will loose
that because the customer gets it for free ...

yes it can suck but IMO just having the 4.5 even for longer won't be
great either. Doing support for so many versions is not cool and really
big deal for regressions: Don't forget: Now there is support for 6.0,
4.7, 4.6, 4.5 BUT in one year there will be support for 6.2, 6.1, 4.7,
4.5 which makes it even worse for backports.

> With the longer release-cycles (up to 4.4), this problem did IMHO not
> exist as every minor version up to 4.3.x was maintained for roughly 3
> years and the overlap was long enough.

yeah but nowadays there is far more code and not more time.

> I'm open to discuss this here, on the phone or also bring in our wishes
> from an agency perspective to the next core-team meeting if we can
> arrange that. As far as I got it from talking to other agencies, we at
> internezzo are not alone with our above mentioned issues with the
> current release-cycle.

Of course you are not alone but on the other side agencies, developers
and everybody tries to get features/bugfixes much faster into the wild.
Maybe you even do a continous deployment and you don't need version
numbers anymore.

Having a longer release cycle would also mean that even more things
change under the hood. Deprecated versions would just be available for
one release instead of 2 because having those is not fun either.

Georg