Skip to main content


Dear #TYPO3 peeps. How likely is it, that you are going to use/ start with v13.0.0 for any serious project?

v13.0 has been released 3 weeks ago and on Packagist, I'm seeing 69 installs (as of now):
https://packagist.org/packages/typo3/cms-core/stats#major/13

v12 LTS has 2.3k installs
https://packagist.org/packages/typo3/cms-core/stats#major/12

v11 LTS has 3.5k installs
https://packagist.org/packages/typo3/cms-core/stats#major/11

v12.0.0 peaked at 134 installs
https://packagist.org/packages/typo3/cms-core/stats#v12.0.0

Wild guess from these numbers: very unlikelyโ€ฆ

wdyt?

reshared this

in reply to Helmut Hummel ๐Ÿ

Iโ€˜m planning to upgrade two projects which are currently in development. But showstopper could be if it takes too long to make all 3rd party extensions and the dev requirements compatible.
in reply to Julian Hofmann

@Xitnelat Depends on the project manager and available extensions. I'd like to start sooner than later.
in reply to Julian Hofmann

@Xitnelat and you are saying those are using latest typo3 versions improportionally more often than composer based typo3 installs?
in reply to Helmut Hummel ๐Ÿ

No, of course not. Should be a joke...
I agree and was surprised by these numbers :-o

In fact, I don't have any new projects at the moment, only upgrades - and v13 is still too young/unfinished to go live in the next few months or the third extensions are still missing (some even for v12).

@helhum

in reply to Helmut Hummel ๐Ÿ

3 answers until now ... 3 answers in2code ๐Ÿ˜‚
This entry was edited (2 months ago)
in reply to MarcusSchwemer

@MarcusSchwemer Letโ€˜s change that. ๐Ÿ˜

I was a bit more conservative in the past. For bigger projects with longer initial development time, I used the latest sprint version. For smaller projects I tended to stick to the latest LTS.

Why? Mostly because of missing 3rd party compatibility (extensions, automating, โ€ฆ).

in reply to Simon Praetorius

@MarcusSchwemer โ€ฆ and I differentiated between small and big projects because ofโ€ฆ money (= risks)
in reply to Helmut Hummel ๐Ÿ

Rather unlikely.

My project with the most experimental approach always has to wait for Georg Ringer's news.

in reply to Helmut Hummel ๐Ÿ

It depends. As always. But breaking changes during the implementation phase are a problem that the customer would then have to pay for. And so I'm actually always waiting for the LTS release.

But I always play around with the current version and see what has changed and what there is looking forward to.

in reply to Helmut Hummel ๐Ÿ

My understanding is that the .0 releases are more useful for extension maintainers than for projects, as projects often need to wait for said extensions to have a compatible release.

While it is cool to hopp onto a new major version early on, it is unlikely to do so for a serious projects that is likely to depend on some existing extensions.

Also, to everyone wondering about the low numbers: those numbers are daily installs and IDK if private packagist installs are included.

in reply to Daniel Goerz

I can also imagine that in such an early version of v13 - if you are already using it - you are more likely to use the main branch instead of a tagged version.

@ervaude @helhum

in reply to Julian Hofmann

@Xitnelat
I think that is even more unlikely, as the main branch is often broken between releases.

I'd not do that. Stick to the releases for projects.

in reply to Daniel Goerz

@ervaude regarding numbers: I don't think absolute numbers matter. But the ratio of LTS installs by non LTS installs is quite meaningful imho
in reply to Helmut Hummel ๐Ÿ

@ervaude And if you only look at the ratio, private packagist installs (which I think still count against public packagist packages afaik), non composer installs or other privately hosted package caches, altogether don't matter as they are evened out.

Unless there is some evidence, that one install type, would prefer either LTS or non LTS versions. I don't think we have such evidence.

in reply to Helmut Hummel ๐Ÿ

Oh, absolutely there is a meaningful difference.

The main point was, that the dot zero release marks the end of breaking changes and therefore enables extension maintainers to tackle compatibility, so that hopefully, a lot of extensions are compatible when finally the LTS is released, so that serious projects can quickly update and do not have to wait for extension dependencies.

I think most projets update from LTS to LTS, wich is reasonable and fine.

in reply to Daniel Goerz

Btw. This is why I think it is a good idea to minimize dependencies to 3rd party extensions.

But every so often this is not feasable, so it is nice that there is enough time for maintainers, at least in theory. :)

This entry was edited (2 months ago)
in reply to Helmut Hummel ๐Ÿ

As Daniel wrote usually not until a majority of extensions picked up. And even there some went v12-compatible not too long ago.
Also your numbers only count composer-based installs. For v13 that might pick up, but I've seen quite a number on v11 or v12 as non-composer.
in reply to Stefan Neufeind

@sneufeind see https://chaos.social/@helhum/111971129756532098 regarding non composer installs...


@ervaude regarding numbers: I don't think absolute numbers matter. But the ratio of LTS installs by non LTS installs is quite meaningful imho

in reply to Helmut Hummel ๐Ÿ

I see very few projects starting out or upgrading quickly to .0 releases (seen a few outliers here and there) but I think the earliest people actually consider might be the feature freeze release, in case too much stuff changes between those few sprints?
in reply to Helmut Hummel ๐Ÿ

I have one private page that I've updated, but 100% of the customer projects won't see anything before 13 LTS.
โ‡ง