Plone 5 will be a brown bag release

Hi there,

I am fully aware that some people will be p***ed by this posting.

The upcoming Plone 5.0 release has major usability issues that I would not expect from a final 5.0 release.

  • with German UI: a mixture of German and English texts
  • ugly typography related issues in the toolbar with improper hyphenation and word wrapping and word truncations
  • completely broken portlet UI
  • many, many tiny but visible UI issues

I documented all found issue in the Plone issue tracker on Github. You just have to search for 'zopyx' reports...you'll see the problems.

I understand that you want to get this release out now - however the quality of this release is not really good enough in order to promote Plone 5 or show it to customers.

Cheers
Andreas

2 Likes

My impression is, that Plone 5 is already quite good, although not perfect. Some issues like the wording (at least the "Sitchwörter vs Stichworte" one) were already present in Plone 4.
One good reason for releasing soon is, that not only you but hopefully more people are contributing issue reports. There is a saying that someone has to wait for a x.1 release, because the first major release is expected to be full of bugs.

However, the expected release date is quite soon - I was surprised by that myself. I agree - annoying bugs can harm Plone's reputation. Another thing we have to care about is marketing. I'd like to see announcements and articles in magazines like iX or Linux Magazin. For that, the release perfection level has to be higher as it is now.

Hi Andreas,

Indeed there are still a lot of bugs in this release, but I disagree with you: I think this release is stable enough and impressive enough to promote Plone 5 and to show it to customers.

I have already started showing it to customers, and they don't say "Ohoh, still a lot of bugs", they say "Waouh, that's really impressive!!".
And that's pretty much the objective when you want to promote something, no?

Cheers,
Eric

@zopyx as a former Zope release manager I hope you never ever made a release without all bugs fixed right? Welcome to software development where one has to balance between features and tiny annoyances that (thanks for reporting them) will be done sooner or later (you are also welcome to fix them as well of course). Actually is not that you are the only one reporting bugs and missing features, there's tons of them. The same can be said for any other software out there, Firefox, Chrome, Linux, all of them have bugs and still all of them keep releasing...

As @thet already says a .1 release is what you have to wait for if you want a more stable release.

IMHO given the amount of benefits it will bring Plone 5 by no means is a brown bag release, it's definitely a major step forward.

Folks, Andreas is right. Even if you do not like it. Cognitive dissonance?

I'd also prefer the release of a series of release candidates instead of a final. I stated this several times. Plone stands for quality and security. While I'am sure its security is fine, quality is missing and a series of RCs (no features, just fixes!) would help to get the final release out fast.

I do not think date-based final releases are a thing a distributed open source project with many small and no major company involved should target at. A date based RC (feature freeze) would work, so forces can be bundled to fix bugs in this phase.

2 Likes

The promotion would be like this: "Shiny but many UI and usability bugs. You have to wait for one more year until the official final release 5.1 is ready". Marketing means that you have to lie at customers.

Andreas

It's all about quality and honesty.

If the locales (as seen in DE) are very incomplete then don't ship the release with German. I can cope with that, customers or potential customers will not. There are so many major and tiny UI defects that in the sum make this release very unpolished. You can discuss yourself if this is the sense of positive marketing for Plone - I have my opinion on that. I need to stress once more on the quality aspect and a missing quality control. I tested the latest beta intensively each time. Usually I could find a dozen very obvious problems without the first 30 minutes (and most of them made it into the bug tracker). I am talking of obvious UI problems. I am sometimes under the impression that testing is solely based on backend test but nobody actually testing the frontend?

Andreas

Zope2 as a backend developer framework is different from a UI-driven project like Plone with the high claims regarding usability and user experience. As I stated in other followup: the current release has many rough edges which appear to users outside the Plone universe as unprofessional, half-baked or whatever.

http://plone5.de/ says (translated): Plone 5 is used in production, new projects should use it now.
-> completely disagree

I am not complaining as an integrator but I see many problems with promoting Plone to other people outside the current Plone universe.

Andreas

I for one am glad for so many good bug reports.
29 bugs that are still open.
Everybody thinking or saying, "go fix em", Please reread this page: https://dev.plone.org/#ContributetoPlone

1 Like

We use Plone 5 for about a year now in several projects. Many issues we found were also fixed by us - including mayor things like linkintegrity or minor things like german translations. I will probably work on the translations again by the end of the week - unless someone else will.
Not everything is fixed or perfectly polished but I consider being able to migrate existing sites from 4 to 5 and from Archetypes to Dexterity says more about the quality of a product than some css-issues.

I'm indifferent.

If we keep waiting, it could take a long time before we release. There are always going to be bug, it's just knowing the right amount to be ready for a release. The experience seems worse for non-english speakers.

I also think, with lots of the bugs being reported as "usability fails" and major problems from Andreas, if he were to objectively look at Plone 4 and create bugs for all the usability problems there, we would have A LOT of bugs for existing final released software. Plone 4, you're used to it because you've been using it forever and Plone 5 it's fresh, different and more obvious.

In any case, the various small artifact problems is troubling. I think a few really good CSS people combing through the site and some translators could squash many of these problems very quickly.

I appreciate how much work Andreas has put into reporting these issue. He's reporting on them because he cares and wants Plone to look good and have a great release.

I'll try to be fixing some every day up to the release day. One thing a tight release date does is force everyone to work hard right up to it. I'm not sure at what point is it too late to delay.

-Nathan

1 Like

Don't get me wrong. I think the UI-issues very important and should really be fixed before the final release. But as @vangeem said most of them should be pretty quick to fix if you know your way around LESS.
So, please designers: Fix them faster than Andreas can discover new ones!

apart from UI issues the most scary thing i found is the complete broken MRO of the dexterity base classes. see my PR here which tries to fix it but invokes side effects. I fear some part in Plone (or its tests?) relies/workaround on the broken MRO

Andreas and the other posters all bring up good points. There was much MUCH discussion over the launch date, and I hope no one has been surprised by the date we chose because we really worked hard to choose and communicate that date among the framework team, various team leads, and the launch team, which had wide representation.

Months ago I asked some key contributors why Plone 5 wasn't done yet, and their response was that no one felt the need to do what needed to be done to get 5 officially released, because anyone who wanted to use it in production already could, if they were comfortable dealing with issues that came up.

So really it comes down to pushing us all to do what needed to be done to get 5 out the door so that mortals can use it too!

I could and likely will write more about what went into getting us all to this stage, which is a very big accomplishment, because we've been hearing how a major and extremely well funded PHP CMS company has had to delay its intended Sept. 15 release indefinitely.

We, however, have been able to get it done, and, yes, there are always bugs, but with your help we know about many of them and we will get them fixed. As a well known release manager has said quite often, people start to take a release seriously once the RC is out...

On marketing, we are doing our best and we will continue to publicize Plone 5. The launch team still has quite a few tasks left on its list!

I totally agree with @zopyx and @jensens.

what starts wrong ends badly: we have committed huge mistakes managing the release of Plone 5 since the beginning (I have talked before about those and I'm not going to repeat myself here), that caused huge delays on the proposed date by doubling the original estimated time to deliver, and now we had no better idea that shoot ourself on the foot by setting a final release date less than one week after the first RC?

who is gonna have time to make a real check on the product quality and fix the bugs? why all of this hurry? we are an open source project, not a public company that have to present "results" to its shareholders.

over the time I have learned a lot with @zopyx comments about how to make good releases, we need to take seriously the quality of our final product: you can't make a RC if you have such a number of unresolved obvious issues; this is not a race.

I understand some of the core developers who had invested a lot of time on this are starting to get tired and frustrated, and want to switch to the next cool thing, but this is not the way we built a name.

please reconsider: let's mark the most important UI and usability issues and lets not make the final release until they are fixed:

Festina lente or, as the great philosopher Dolly Parton used to say, I'm gonna hurry just, just as slow as I can…

1 Like

There's a major misconception here that needs to be addressed: that my cutting a final release is dictated by press releases. Will we make it by next Tuesday? That'd be great. Will I push on the 15th if I don't feel it's ready? No. If it takes ten more pre-releases to get everything to a state I'm happy with, then we will have ten more pre-releases. But we are now days, not weeks or months, away from Plone 5.0 final.

Andreas' complaints are exactly why we're at RC state though. It has always been the case that the average Plone user doesn't install a pre-release until Plone hits RC status, while only the more seasoned users try the betas. So, with this release, we've just opened Plone 5 up to the people better attuned to these sorts of user-facing bugs. And now that the feature fixes are out of the way, developers can focus on clearing those up.

To prove my point: in the 12 hours since the release, I've seen 13 new tickets, two translation fixes, a number of bug fixes (including some of Andreas' listed items), two package releases, and four pull requests. This is a planned process, not a race.

So, Andreas, thank you for complaining in ticket form. It's making us better. I'd encourage everyone to do the same.

4 Likes

thank you, @esteele! that's why you're our beloved release manager.

unfortunately all of this came on a time we have way too much work over here; anyway, next Friday is Plone 5 day at Simples Consultoria; we'll be evaluating the RC and helping on reporting and fixing issues.

I just looked through the issues entered by Andreas in the last day or so. These are great tickets! Resolving them will improve Plone 5.

What these tickets are not, though, is "blockers." It's not as if work will stop with the release of 5.0. Tickets like these tell us we'll need a 5.0.1 soon and a 5.0.2 soon after. Release early, release often, right?

I don't think anyone is pressuring Eric to release if there are blockers. In fact, one of the reasons why we can talk about a release schedule is that we know he'll stand up to anyone if needed if P5 isn't ready.

1 Like

I don't agree, @smcmahon; there are many issues with the toolbar, for instance, that need to be addressed before a final release:

Since many years I am a user of Plone, since version 2. I think it was back in 2003. I once installed Plone 4.1 with more than 100 add-on components. The quality of Plone out-of-the-box was allways very high. I had more issues with the add-ons...

I love the evolution to Plone 5 with the smashing new GUI. I just installed 5.0rc1 and found it really nice. Yes some nasty GUI bugs are visible and should not be there as it gives a false picture about the quality. The number of bugs listen here however gives me the impression that the current release date is set too early.

This might be true or false. I cannot judge that properly. I do understand that Andreas is really worried. When you release version 5.0 then please make sure that the most obvious bugs have been solved. And make sure that version 5.0.1 is planned shortly after version 5.0. Get it fast to a high quality level, which we all expect from a product like Plone...

Additionally: yesterday 11-9-2015 I installed Ploneintranet development version which comes with Plone 5.0b4 and I replaced that with Plone 5.0rc1. All worked fine. I got only 1 bug in the test suite. It shows that Plone 5 is allready very well on its way.

I want to give compliments for both teams: the Plone 5 team and the Ploneintranet team. Two wonderfull solutions.

I am working for Europe's largest IT company and recently introduced Plone 4.3 which is now very succesfully used as an Intranet solution including PloneBooking (Yes I managed to get that working properly!). I use a setup with HAproxy. Ningx and Varnish (Solr is used as option and further some other nice add-ons like the Document Viewer).

I use Plone 4.3 also for a very small student start-up company in Ethiopia (vpelectricalpower.com. This is running on 4MB memory on a VM with Ubuntu in the cloud including Zimbra and Piwik ,HAproxy. Ningx and Varnish. OpenLDAP... Yes I know all very minimal, but enough for a start-up with a very small walllet... however all is working fine for them and I didit for free for them :-).

I have evaluated Joomla, Drupal and plenty of other open source (CMS) systems the last 15 years, but it is really Plone that offers highest quality with an extreme flexibility to integrate it with other solutions.

Therefore: Thank you and well done!!. Keep the quality high, but excellerate as well as competition is stiff. I do have great expectations from the Ploneintranet solution and of course from Plone 5!

4 Likes

Plone Foundation Code of Conduct