Lessons from a Botched Upgrade

[frommagio]

Now that 2+2 is operational again, I thought I'd post a few technology 101 thoughts. I'm only scratching the surface here; these are just the ones that occur to me quickly as I type them in. Believe me, I'm probably going to miss 25 rookie mistakes in this quick list.

There are lessons to be learned, from a business and systems I/T perspective:
- Choose your technical staff wisely.
-- You know your business, techies know technology.
-- Many techies will be able to understand your business.
-- It's your job to find one.
-- Remember that your techie works for you.
- Consider your reasons for upgrading.
-- What are you trying to accomplish? What are you risking?
-- What are the drawbacks to upgrading? The benefits?
-- What if you decided just to stand still? Any problems?
-- Is this for business reasons, or for your techie?
-- If it's your techie, does he understand your business?
- Plan your upgrade.
- Test your upgrade first
-- Your customers are not a testbed.
-- Use a separate staging area.
-- Use artificial load generation to simulate real life.
-- Be sure that the new features are working as you hope.
-- Remember to check that the basic stuff still works.
-- Make a testbed beta available to some helpful customers.
-- Review the planned benefits, every point. All proven?
-- Review the risks, every point. Satisfied?
-- If your techie didn't do all of this, you screwed up; bad hire, learn the lesson.
- If things look good, announce your planned upgrade
-- Give plenty of warning to your customers
-- Give yourself plenty of time to back out
-- Remember that the software works for you, not vice-versa
- Finish your testing
-- Don't hurry.
-- Don't be afraid to back out; believe your own eyes.
-- Accept a schedule slip if it's too early to decide.
-- Remember that you are in charge.
- Go for it
-- After all your worries are satisfied, pull the trigger.

[john voight]

Lesson: Go with Vbulletin

[Shoe]

Also, once you become a major site, it is usually a bad idea to just completely throw away what is working for you and come out with a whole new look and feel. You became successful for a reason. You don't see Yahoo or Google completely redesigning their sites (other than gradual changes over time). Even when Yahoo changed the look of their homepage a few months back, the same basic feel was still there.

This is a forum. One of the reasons it is better than 99% of the forums out there is because it is not like 99% of the other forums out there. This forum (in its current state) is fast, easy to use, and easy to navigate. True, there are a few things that could be added or changed (such as the green bar on the left that has out-grown its usefulness), but for the most part, everything works as perfect as one could expect.

[Ryan Beal]

We did most of this.

Regardless, it looks like we won't be seeing UBB 7 again. I'm a fan of vBulletin too, but we obviously haven't made a decision about which way to go in the future yet. I look forward to reading more from you guys about which board software you like.

[tuq]

[ QUOTE ]
Regardless, it looks like we won't be seeing UBB 7 again. I'm a fan of vBulletin too, but we obviously haven't made a decision about which way to go in the future yet. I look forward to reading more from you guys about which board software you like.

[/ QUOTE ]
I have no opinion, but it seems like about 70+% of the bitching involved the speed of the forums. If they whizzed along or otherwise weren't perceptibly slower than previously, the complaints would be nearly minimal.

That seems to be the biggest gripe, by far, and I have no idea how it's tied to the forum software. To me they would be two totally different, unrelated things, but apparently not.

[StevieG]

[ QUOTE ]
I have no opinion, but it seems like about 70+% of the bitching involved the speed of the forums. If they whizzed along or otherwise weren't perceptibly slower than previously, the complaints would be nearly minimal.


[/ QUOTE ]

There were still some legitimate usability concerns, but yes, tuq, the speed was the real problem. I agree with Jack that these other problems could have been solved in short order (though not his insistence on labeling them cosmetic).

However, the performance issues made the entire situation untenable.

[Madtown]

[ QUOTE ]
[ QUOTE ]
Regardless, it looks like we won't be seeing UBB 7 again. I'm a fan of vBulletin too, but we obviously haven't made a decision about which way to go in the future yet. I look forward to reading more from you guys about which board software you like.

[/ QUOTE ]
I have no opinion, but it seems like about 70+% of the bitching involved the speed of the forums. If they whizzed along or otherwise weren't perceptibly slower than previously, the complaints would be nearly minimal.

That seems to be the biggest gripe, by far, and I have no idea how it's tied to the forum software. To me they would be two totally different, unrelated things, but apparently not.

[/ QUOTE ]

They can be quite related.

Bring on the vB. Please, love of God, oh please. And do some research into the design principles used by some of the larger forums around.

While I was no fan of the design used for the upgrade, I hope that a lot of criticisms would have been taken into consideration, and the design slowly improved. Hopefully more people will be on the lookout for the test area for the next upgrade, so that more feedback is given before release. But the main point is that UBB isn't suitable for a forum of 2p2's size and hasn't been for a long time.

The main concern I have is that conversion from one forum software to another can be tricky and painful -- look at the PM issue the upgrade had, and that was a UBB upgrade. Hopefully vB won't result in those kinds of issues. I'd be perfectly fine with just starting with a fresh install of vB, and archiving everything posted on the UBB servers. But many people wouldn't, so conversion will be a delicate matter to handle.

[StevieG]

[ QUOTE ]
I'd be perfectly fine with just starting with a fresh install of vB, and archiving everything posted on the UBB servers. But many people wouldn't, so conversion will be a delicate matter to handle.

[/ QUOTE ]

My knee jerk reaction to this was that this would be a bad for 2+2 from a business standpoint.

But would it be?

I reference old posts often, but not many do. The forum's real value comes from the size and speed of the community. As long as login information were propagated, would it really matter?

[suzzer99]

I've been doing web programming for 10 years. At my last job I was the chief UI designer leading a team of people on a $5 mil project. The best user interface always seems completely straightforward and obvious in retrospect. But getting to that point is nowhere near as easy as people think. IME it always takes a ton of hard work and a certain kind of fairly uncommon puzzle-solving-ish ability (yay).

There was nothing unfixable from a UI/aesthetic POV in this latest upgrade from what I saw, and actually some promising new features. Unlike the previous "woodie" upgrade, which was unsalvageable. But how anyone could think that was ready for primetime is beyond me. Please consult a professional next time.

I'll cut you a deal. [img]/images/graemlins/wink.gif[/img]

[Madtown]

[ QUOTE ]
[ QUOTE ]
I'd be perfectly fine with just starting with a fresh install of vB, and archiving everything posted on the UBB servers. But many people wouldn't, so conversion will be a delicate matter to handle.

[/ QUOTE ]

My knee jerk reaction to this was that this would be a bad for 2+2 from a business standpoint.

But would it be?

I reference old posts often, but not many do. The forum's real value comes from the size and speed of the community. As long as login information were propagated, would it really matter?

[/ QUOTE ]

Well, it would be a horrible idea. It should never be done. I'm just saying that I personally wouldn't care, as long as everything was archived.