Posted By: jeffsand | Jun 6th @ 6:50 PM
page 1 of 2
Comments: 43 | Views: 887
It has been a hell of a week. Lots of work on our end to get V4 of 9 stabilized in production and lots of patience on your end while we fixed things that were broken. 

Thank you for bearing with us as we made this deployment and for being so patient we deployed V4 first to all the other communities. We still things to fix and clean up, but we believe we're on a steady path of improvement.If you see something you want improved, have suggestion on how to change the UI text or find a bug, please report it on the Wiki.

The great news, is that with this week behind us and V4 now running on all of our sites, we can evolve so much more rapidly. You'll once again see constant improvements on 9, that the early Niners saw..

Thanks everyone again, you are why we do what we do.
-Jeff
I met Lenn six weeks ago and told him how I thought C9 had become part of Microsoft PR.  He understood what I was talking about.
Now if only my account would work...

But overall - fantastic job. The perf on this site really destroy's the old C9 and most boards I frequent on the net.

Design looks pretty polished. Clean.

Well done.
Now I get why people don't like you, eagle.
Lenn was very happy to meet me, don't you have anything nice to say about the new C9?
It really is fast now...  great job C9 team!

Looking forward to seeing what you can cook up now that you've got a codebase that's actually workable...  hopefully you'll get feature parity soon () then wow us with some never-before-seen internet forum awesomeness (I have no idea what that means, but it sounded cool).
I've been impressed with how the C9 team was so responsive to feedback when v4 went live. Very impressive. Congratulations on shipping v4, I hope you guys had a huge party or something. You deserve a pat on the back. =)

It's cool how the stuff works now... A lot of the features are still missing, but we got the forums back and the videos are now also available in more formats. Great work team, keep it up and give us a lot more of the c9 v4 goodies.

Just wanted to say that you've done a great job here guys.  My hat is off to all of you Smiley

I know I've not been here all that long, but I've enjoyed myself here so far, and the sense of community is brilliant.  I used to be a part of the linux community, but got tired of the snobs and getting told to google stuff I didn't understand....

And once again, thanks.
I would like to know why was it so important to roll out a site so lacking in functionality (and testing) like quoting & going directly to last post & embedding video & etc...

(I'm really "replying to thread" here)
Because the communities are a showcase for Microsoft and it's technologies?

I like the new C9!
Thanks for the comment Minh... there actually is some reasoning behind it. We consider Channel 9 to be our main or flagship property, and we have a lot of plans for it beyond what you have seen her already. Some of those plans required new features, and some just required more flexibility. It was decided that it would be difficult to build anything new on top of the code that was already running.

Obviously there are some features missing, but we have been working on shipping this site for nearly two years now, and at some point you have to start cutting items off the feature list so that you can ship, and we decided that the core functionality of the site is based around three things: watching videos, commenting on videos and posting in the forums. There are, of course, more things than those three going on with Channel 9, but those three were the ones that we thought we couldn't do without... not even for a few weeks.

None of the missing features that you mention (quoting, going directly to a last post, embedding video) prevented that basic functionality from working, so instead of waiting even longer and preventing us from working on the various new features that we hope to add to Channel 9... we decided to ship once we believed the basic functionality was complete.

Some of the bugs that turned up, didn't turn up when we tested... we did run load tests against the code, in addition to using this exact code in production on several other sites... and it turns out that nearly all of the problems that people saw in the first week after we shipped were tied to performance, and performance was being severely impacted by ongoing app restarts (nearly every couple of minutes). These app restarts didn't happen when we tested, and they didn't happen on the other sites, so we never ran into the rest of the bugs that these restarts ended up producing. Why is that? Well, it is hard to say... we did load testing with an exact set of requests from a recent day on the old channel 9, so the usage pattern should have been the same, but it turns out that in real usage ASP.NET was continually trying to recompile all of our pages, and after 15 recompiles it would automatically recycle the app... all unexpected. And it took us a day or two to find that problem. Anyway, obviously our testing was not complete and comprehensive, but it wasn't from a lack of importance placed on testing, it was merely a mistake that led to some very poor results after roll out.

So... why was it important to ship out Channel 9 when we did, instead of postponing that ship again and again until we reached some mythical level of confidence that it was ready? Our confidence wasn't going to get much higher, and we felt that any issues that turned up would happen quickly and would hopefully be addressed quickly as well. To a certain degree, we also felt that we needed real users to use and comment on the site features... and that we couldn't get that level of participation by just extending the beta period longer. None of us were pleased with how those first few hours and even days went with this release, but that was after waiting two years after we planned to ship C9v4, I doubt waiting another month or two would have made much of a difference. Does that mean that you, the unfortunate user, had to unwillingly help us test... yes, and that is why Jeff is thanking you... for your patience and understanding.

I'd like to thank the Channel 9 users as well, because many of them were very patient and understanding, but I also think that in your case Minh, there is little we could have done that would have left you pleased. If we had shipped with those features included, you would have found something else to complain about... which is fine, every group needs that person who looks for something to complain about, that person who seems continually unhappy... and you seem to have appointed yourself to that role for Channel 9.

We are working hard to address the great feedback we have received over the past days, and it will result in many changes including some new features, but I think that if our goal was to make sure that you were happy with the site, then there would be little point in continuing at all.
I think that the Niners are very forgiving about the initial problems, because we are developers ourselves and we know exactly how you must feel. How many professional developers here have experienced working on a project for ages and minutes after it's deployed, it doesn't behave like you wanted, sometimes even crashing? Like Maddus Mattus' avatar demonstrates, "It works on my machine" has become a common phrase that many developers will recognize.
Duncanma wrote
Obviously there are some features missing, but we have been working on shipping this site for nearly two years now, and at some point you have to start cutting items off the feature list so that you can ship
OK, I read ahead and you promised not to be defensive. Look. If I were to come to my boss & say, what we have is a very poorly construct app, I can rewrite it so it'll be easier to maintain, add new functionality to. Oh, and BTW, I'd have to cut out 50% of the existing functionalities, I'd be laughed at really hard. So, you had your plan ass-backward. You added the new stuff at the expense of the old stuff. What I would have done when talking to my boss is that, I had to promise to implement 100% of the existing functionalities at the expense of no new stuff coming out for a while, but at the end of the journey, we wouldn't lose any functionalities, but having a nice codebase for the future.


Duncanma wrote
that the core functionality of the site is based around three things: watching videos, commenting on videos and posting in the forums.
The order of those priorites are very apparent now.


Duncanma wrote
There are, of course, more things than those three going on with Channel 9, but those three were the ones that we thought we couldn't do without... not even for a few weeks.
And in the process you broke old posts (where there were embedded videos), you broke the wiki (old wiki formatting codes don't translate 100%), but hey, we can watch the UGTV episode, so that's all right.


Duncanma wrote
None of the missing features that you mention (quoting, going directly to a last post, embedding video) prevented that basic functionality from working, so instead of waiting even longer and preventing us from working on the various new features that we hope to add to Channel 9... we decided to ship once we believed the basic functionality was complete.
Well that's a judgement call, isn't it. And it's a bad call.


Duncanma wrote
There Some of the bugs that turned up, didn't turn up when we tested... we did run load tests against the code, in addition to using this exact code in production on several other sites... and it turns out that nearly all of the problems that people saw in the first week after we shipped were tied to performance, and performance was being severely impacted by ongoing app restarts (nearly every couple of minutes).
Well, we're all developers here. Perhaps a postmortem of why you didn't see the bugs in testing. I'm sure someone will benefit from it.


Duncanma wrote
So... why was it important to ship out Channel 9 when we did, instead of postponing that ship again and again until we reached some mythical level of confidence that it was ready?
There's a much more practical metric than the "mythical level of confidence". It's "percentage of pre-existing functionalities implemented". Why is it that you didn't think of that?


Duncanma wrote
I'd like to thank the Channel 9 users as well, because many of them were very patient and understanding, but I also think that in your case Minh, there is little we could have done that would have left you pleased. If we had shipped with those features included, you would have found something else to complain about...
I guess we'll never know if I'd be please or not. The fact is you shipped a product lacking pre-existing functionalities. Your launch was horrible. I still am getting hung request today.


Duncanma wrote
which is fine, every group needs that person who looks for something to complain about, that person who seems continually unhappy... and you seem to have appointed yourself to that role for Channel 9.
Oh my goodness. Now, I'm a complainer because I pointed out the obvious. You can make this about me & not about your plan to implement a broken v4. What do you want me to do? Smile & say, "good job, Duncan." You want me to pat you on the head? It was a FUBAR'd roll-out. And the site is much better now. What I'm complaining is that your plan for v4 from the start has MISSING pre-existing functionalities that I use ALOT. That just wouldn't fly in the real world.


Duncanma wrote
We are working hard to address the great feedback we have received over the past days, and it will result in many changes including some new features, but I think that if our goal was to make sure that
you were happy with the site, then there would be little point in continuing at all.
I read through the suggestion list, and it contains a lot of what I'm saying. MISSING FUNCTIONALITIES. I'm sorry that I sound harsh. That's just my personality. But you have to man up to your mistake of planning & delivery the new version.
I don't care about the initial roll-out problems either. But what still amazes me is the MISSING FUNCTIONALITIES. And not implementing the is actually in the release plan.
If it was all about showcasing MS tech, then we have a problem. Tech can't be tech for its own sake.  C9 forgets the motto "Don't piss off your customers."
I have to say I agree with Minh. I think the new C9 has great potential, but I would never have released it in the state it's in now. I would've waited until feature parity with C9v3 was accomplished.
offtopic: I find it interesting that we don't have a SPAM button and that nobody spammed so far... kind of strange - i wonder why.

The new posting mechanism is ajax based with no non-javascript fallback. As far as I know spambots haven't figured out how to do that yet.

True, but still there is no real dude doing anything... let's hope it stays so Smiley
Yet here we all are, still posting away.
You can't deny that before the launch of C9v4, people were comparing C9v4 with Duke Nukem Forever, claiming it was vaporware and would never get launched. There was pressure from the community to launch the damn thing. The C9 team says they want to be sure that the core features are implemented before they launch it. People complain that the team first applies Evnet to other sites and that C9 is ignored. The C9 team explains that C9 requires a lot more features and that the other sites are good tests for the platform. People keep complaining, some say that it might be better to just launch it and add features incrementally. The C9 team launches C9v4. People complain that features are missing. They can never satisfy everyone. The team made a decision (a pretty brave one) and I believe it was the right decision.