CG12 Keynote Video

Wednesday, June 13, 2012 by Peter Gregory

Watch the CG12 keynote video where Niels explains the reasons why it was decided that development on v5 could not continue.

26 comment(s) for “CG12 Keynote Video”

  1. Gravatar ImageBilly Koch Says:

    Excellent Keynote! Thanks to you and everyone who have all contributed to the Umbraco Project you guys ROCK!

  2. Gravatar ImageMichiel Says:

    Here is the thread referred to in this keynote: http://our.umbraco.org/forum/core/umbraco-5-general-discussion/28565-Umbraco-5-Performance-issues

    Niels,

    I can definitely understand what you are saying here. From the outside Alex's comments sure sound like a lot of 'highfalutin engineer' talk, and overall it's not very... inclusive. (Though I highly respect Alex's work, I think it's amazing!)

    I went and read the whole thread just now. I think that thread is fascinating, informative, hopeful, (mostly) optimistic and overall shows a willingness from everyone involved to get things right. At one point, Niels, you commented that it was 'painful' and 'heartbreaking', but that was really the only sour note in the thread up until that point.

    In fact, later in that same thread you yourself confirmed multiple times your dedication to improve v5 ("I know we will succeed, but it will take time.").

    So I really can't see how from there you would jump to the conclusion to drop v5. Especially because in the beginning of the thread I sense that (despite performance problems) people are genuinely enjoying building websites on v5. It would seem that the community has (or had) an overall more positive outlook.

    Remember the good old times when v5 was the future, and performance was just another issue to be tackled? http://umbraco.com/follow-us/blog-archive/2012/1/4/umbraco-5-on-performance-and-the-perils-of-premature-optimisation.aspx

    I say: let there be forks!

  3. Gravatar ImageTroels Says:

    You´re not the first community to have issues with the v5´s. Seems to be a lot of redundant experiences and considerations shared with what the TYPO3 community experienced and in the end have come to a similar conclusion and forward going strategy (It took them years to come to the same final conclusion, though, so acknowledgements from me for making a resolute decision at an early stage saving a lot of peoples efforts and money)

    Read this string: http://buzz.typo3.org/people/xavier-perseguers/article/typo3-60-at-the-corner-how-is-it-possible/

  4. Gravatar ImageMatias Wald Says:

    Don't accept this! we need to ask Umbraco to release all assets for V5 to the public! Let's make this a true community driven open source project!

    If you support this idea, I've created a petition online so Umbraco HQ takes notice!

    http://www.change.org/petitions/umbraco-hq-release-all-assets-for-umbraco-v5-to-the-public-domain

    Thanks in advance!

  5. Gravatar ImageHartvig Says:

    @matias no need for a petition, v5 is MIT and everything is out there in the public on Codeplex and always have been.

  6. Gravatar ImageHartvig Says:

    @michiel We've been analyzing for weeks and we've been 18 people gathered for three intense days to come to a conclusion. None - zero - of us thought there was any merit at all to keep trying to fix v5 as we all agree that the core of it is fundamentally flawed. But as we've said there's big parts of v5 that can and will be moved to an improved v4 in the very near future. Including the MVC and plugin engine parts.

  7. Gravatar ImageMatias Wald Says:

    @hartvig Are all assets public domain? I don't think so... what about architecture documents, testing plans, use cases, sequence diagrams, etc? What about "why" is V5 fundamentally flawed? (disclose the analysis you and your team made)

    "V5 isn't a natural succesor to V4" (from keynote). Yes, V5 IS a disruptive change, but one long awaited and very natural considering MVC3 (and MVC4 in the horizon) and the way web development will be done in the future. Is V5 ahead of its time? for sure! Should we kill the project because some components used by Umbraco are not yet up to par? Of course not. If something isn't working you either fix it or remove that feature for now (or you build from scratch which is not an option at this point).

    I say no to killing "Umbraco V5", disclose and publish every asset related to this project and call it "Open Umbraco V1"!!

  8. Gravatar ImageHeather Floyd Says:

    To Niels & the core team - what a great presentation. It takes guts to admit mistakes and trash a lot of work, and I'm sure it wasn't easy. You're going to take a lot of flak for awhile over this, but I feel very optimistic about the future of umbraco and am still devoted to the (v4) platform. I am especially pleased to hear about your efforts to open up the development process, it's been my experience that good ideas become extraordinary when exposed to the light of others feedback.

  9. Gravatar ImageSjors Pals Says:

    One question remains, why was V5 released as stable if the core team knew that there where such big flaws in i?

  10. Gravatar ImageMantorok Says:

    @Sjors

    That's a question that is on the end of everyones lips right now, and I would very much like to know that, how could such a beast get out the door into the public?

    Confidence will always be an issue until that question gets an answer IMO.

  11. Gravatar ImageGeorge Says:

    Blaming the community for being silent/lazy is quite rude. It is a natural reaction from a decision of not including the community in a development process of v5.

    Being honest doesn't cover the fact that you released bad product before it is tested.

    Great architecture doesn't cause such problems where fixing of one causes few more.

  12. Gravatar ImageMichiel Says:

    I'd be very interested in the analysis that lead to the conclusion that the codebase is _fundamentally_ flawed. Do you mean that at a technical level (code, architecture), or at the level of interaction with the code (API, concepts)?

  13. Gravatar ImageChristian Foged Says:

    One thing is to release a flawed product that you have been working on (and hyped quite a bit) for 3 years.
    Another, and far more difficult thing, is admitting you were wrong, so early on after the initial release.

    Have to say that is sucks a lot that V5 ended up in the can, but it was for the right reasons it ended up there.

    Kudos to Team Umbraco for showing leadership, learning from their mistake, and taking the Umbraco project back to basics. Loved the keynote, honesty is the way to go... Looking very much forward to 4.8 btw...

    Going forward I personally hope that the trust level of Umbraco won't take too much of a beating, especially for those having invested a lot of time in V5.

  14. Gravatar ImageSebastian Says:

    Extremely bad news! One of the big reasons we got started with umbraco was MVC and all new techniques in .net. I thought the project looked so promising. I dont understand the descision to kill it.

  15. Gravatar ImageRadek Rezac Says:

    I had made decision between Umbraco and Orchard, because I want MVC. I decided fo Umbraco and Iwas looking forward for v5. Now It is clear: Orchard. Bye Umbraco.

  16. Gravatar ImageJohn Says:

    @Radek I'm quite interested in your view of orchard vs umbraco and your views of leaving umbraco completely.

  17. Gravatar ImageAndy Hawken Says:

    I'm in a similar position to @Radek I'm an MVC developer. I also picked v5 as the tool of choice over Orchard - a decision discussed on stackoverflow http://stackoverflow.com/questions/5565086/is-orchard-or-umbraco-mvc/11024913#11024913 . Most of the (non CMS) code I write is on MVC projects so I want an MVC based CMS. v4, even with Razor, isn't MVC. Orchard is. So that pushes me back towards Orchard. (Sadly there is a lot of Orchard architecture that I don't like too so unlike @Radek I don't have a clear decision yet).

    I found v5 really easy to work with, SurfaceControllers really useful to write and easy, and of course with Razor as a templating engine I could do so much.

    Full disclosure - I was one of the day 0 5* reviews - my 5*'s were based on building a farily complex beta of a site for a client from scratch in a couple of days based on the nightlys. The only issues I had were performance, and I was happy that would get sorted. (cough)

    So the question now is whether I learn 4 (I have zero experience) or go back to Orchard. Don't really know yet.

  18. Gravatar ImageEnome Says:

    It's funny how the keynote blamed the community for not being critical enough. Did you ever try saying something bad about Umbraco on Twitter? I am all for defending your product but getting 10 replies in a matter of minutes when you say something bad makes me feel a little: http://blog.tmcnet.com/blog/tom-keating/images/simpsons-mob-torches.jpg

    I stopped using Umbraco at 4.7 mainly because of performance issues. I tried to ask about it on the forum but I never really got a good answer. So I concluded that Umbraco needed a lot of memory to run. You could argue that Umbraco is an enterprise product and the sites I made were mostly for local businesses with limited resources. I think most of my sites have like an average of 50 nodes. To run these comfortably I would need to have around 300MB ram for each website otherwise my app pool would recycle every few minutes. I always had a hard time explaining it to my customers why someone with a 1 euro a month hosting running some shitty php website could out perform Umbraco. Also if you can't run small sites how do you scale up from there?

    After Umbraco I tried a lot of other open source projects. I now have experience with RoR, Django and lately Express.js. If you compare the quality of these open source projects with Umbraco it's kinda puts everything in perspective. I guess having to create an open source project in such an aggressive Microsoft environment with the newest best features lurking around each corner can be a little intimidating.

    Anyway take this rant with a grain of salt since I am not an active Umbraco user. I still have some frustrations thought because Umbraco really had potentional and the community is great.

  19. Gravatar Imageesunxray Says:

    @Radek Rezac
    Just like you, I want to go V5 from V4 because V5 is based on MVC. Now, I must make a new choice to select a CMS based MVC. Orchard is not so easy for end user, because it needs to code. I think Orchard is focus users who are able to develope fuction. Then, I select another CMS named Kooboo. I like it just because it is so easy to use. Some front end code just modify exam;le codes which have included in final released version.

  20. Gravatar ImageGijs van Dam Says:

    It shows courage to discontinue V5, and the keynote addresses that in a way that commands respect. But the keynote doesn't address at all the reasons why dev on v5 went so totally out of sync with what Umbraco stands for.

    How could this happen?

    And more importantly what measures are taken to prevent it from happening again?
    If there's something fundamentally wrong with the product v5, than there's something fundamentally wrong with the team that created the product in the first place.
    Saying that the community should speak up sooner, or be more critical is a huge sign that the v5 team and Niels still have a blind spot for what's inherently wrong in their own processes. It's a sign of looking at others while you should be looking at yourself.

    And although I respect the courage that it takes to be on the stage and give this keynote, I have to say that the keynote lacks substance. Because if it didn't, the issues of why and how did we get here would have been properly addressed.

  21. Gravatar ImagePeter Bailey Says:

    Ayende has posted an overview of Umbraco V5 and NHibernate which might help people understand the performance issues, and the architecture that led to it: http://ayende.com/blog/156577/on-umbracorsquo-s-nhibernatersquo-s-pullout (question originally posed by Michiel)

  22. Gravatar ImageEric Herlitz Says:

    I’ve been developing stuff since the late 80’s and have been into web development since it started. MVC is a truly overvalued technique that actually doesn’t always help you. I have been in several MVC projects, not only in .net but also in php and some other languages.

    MVC can be great for modules, it can be great for some parts of a business layer, it can be…

    If I had been one of the 15 people invited to the plans three years ago I would have told them this. Question is if they would have listened to me, probably not. I expect a lot of angry comments for degrading MVC, thing is that I’m not degrading it; I simply try to be realistic, based on experience. I can handle MVC, but have been developing for all my life. Umbraco and web development in general isn’t there yet and probably won’t be for a long time.

    Back to the 15 people, shame on you for not analyzing the audience but instead playing your own game.

  23. Gravatar ImageIndrakumar Says:

    I was so excited to start using version 5 but got to hear a sad news. I hope we all together can make our Umraco CMS a great CMS ever. I Love Umbraco and will continue to Love it.

  24. Gravatar ImageJohn Says:

    It is a brave step and certainly the right one. What really bugs me is that Niels very often complains why nobody warned him. Reminds me of this great scene:
    http://www.codinghorror.com/blog/2012/07/but-you-did-not-persuade-me.html

    Let me quote Joshua Bloch from his Google Talk How To Design A Good API (2007):
    The worst thing you can do is send six smart guys into a room:
    http://www.youtube.com/watch?v=heh4OeB9A-c&t=10m05s

  25. Gravatar ImageHartvig Says:

    @John: It's been mentioned a couple of times and it might have came out wrong / lost in translation. I wasn't trying to move the responsibility from me to anyone else - my point was that a community who gives honest - and critical - feedback is better than a community of yes-men.

  26. Gravatar ImageAnthony Candaele Says:

    I think it's fair to say, almost 7 months after the CG12 keynote, that Umbraco survived the V5 crisis, and actually became stronger of it. Admitting the mistake and stopping the development of V5 was without a doubt the best choice to make.
    The participation of the community in the development in Umbraco has never been bigger, MVC is integrated in Umbraco, and a lot of exciting developments (Belle, Concorde, V6) are in the pipeline.
    It really proves that Umbraco is a framework by and for the community, and that it is resilient to crisis.

    A big cheer for Niels, the Umbraco Team and it's amazing community for sticking together in good and bad times.

    I whish the whole Umbraco community fun times in 2013 and I'm already looking forward to CG13

    Anthony

    Anthony

Leave a comment