Connect Digest : 2011-03-12
Last year, I came to a very tough decision that I would cease publicizing Connect items in an attempt to drive up votes and get important issues fixed. This was almost entirely due to a couple of MVPs criticizing me for raising awareness of certain Connect items instead of letting them be found "naturally." I wasn't sure what world they were living in, where droves of everyday end users just happened to stumble upon Connect items without any prompting. I suppose it could be said that the more users that come across a bug, the higher the vote count will be – however history has shown that it is much more likely that there will just be more copies of the same bug, since folks will report the bug on their own after a half-hearted search for a previously existing submission. All this does is dilute the votes and make it harder for Microsoft to assess the community's opinion about the bug.
Suggestions are a different beast altogether. It is quite often the case that I see and up-vote suggestions that I would have never thought up on my own, but once I read the submitter's point of view, am swayed to vote with them (or against them, in some cases). Is this the same experience for a lot of people? I doubt it is extremely common – I'm a Connect power user and actively monitor the RSS feeds for both new feedback and recently modified feedback for SQL Server. It is my belief that, without a little nudge from folks like me, suggestions and bugs would go unreported in the first place, or would wither away due to little Connect activity. Most people have better things to do with their time than trawl Connect to stumble upon new items they agree or disagree with.
At the MVP Summit last week, I was encouraged by several of my esteemed peers to resume my Connect Digests, and to not listen to folks who think they're a bad idea. The overall benefit – getting Microsoft to see the community support behind a particular idea – far outweighs any negative they could envision.
In addition to this, over the last few days, I've had several of my suggestions dismissed as "won't fix" by the product team. Each closure invites me to re-open if I disagree; however – as a few other MVPs pointed out (since this also happened to them recently) – the submitter can't re-open suggestions, only bugs. And if you submit a new suggestion in an effort to gain new visibility for your idea, unfortunately you lose all the votes from the closed item (except your own, of course). This made me feel that in order to help prevent suggestions from being summarily dismissed, they could use some more community attention… though I confess there were 20 or more votes on several of the items I've watched being closed. Still, more votes can't hurt.
So, I am going to ignore the cries of foul play from a few noisy individuals, and resume publicizing some Connect issues where and when I see fit. If you don't like it, I'd love to hear your reasoning, but I have my suspicions about the underlying motivation, so I'll take it with a grain of salt.
So please vote…
This past week, there were two really important issues brought up on the Connect site, and I hope that Microsoft can be convinced to properly investigate both.
Daniel Schmitz reported something that many users are complaining about: Visual Studio 2010 SP1, released earlier this week, is breaking their IntelliSense in various ways. I've seen reports about both SSMS 2008 and SSMS 2008 R2, and I've seen hints that the bug's manifestation may or may not occur on Windows 7 or Windows 2008 R2 if you have installed Service Pack 1 *for the O/S* first. If you can (or can't!) reproduce this problem, please post to the Connect item (or comment here) with your operating system, whether or not you have SP1 enabled if it is Win7/2008 R2, the exact SSMS version (help | about), and your actual observed symptoms. Right now the information is a bit of a hodge-podge, and in order to get the issue resolved, it will likely take a much larger sample of complete information.
Fellow MVP Jonathan Kehayias recently suggested that Extended Events could include actual execution plans as part of the collection process, to greatly enhance performance troubleshooting efforts. As most of us know, when we collect showplan information in a trace, it is great to see the plans but we can't figure out what statement or object is associated with the plan. Why? Because TextData is always empty in the trace. You have to also be collecting statement information in your trace, and then you must manually correlate the statements with their plan counterparts. As you can guess, rigging Extended Events the way Jonathan suggests will be much tidier.
In a lot of cases, Microsoft doesn't just want to see vote count; they want to see your opinions about WHY you are up-voting. They need more than just a raw number to show their bosses, and for them to even be behind the idea in the first place, they may need more business/use case behind the need for the change. The original item may not have been descriptive enough for them to justify the work required for the change, and you may have some insight that will win them over.
And if you are down-voting an item, please consider telling the author why. I've had down-votes against issues where there was a typo in Books Online – I don't know who would vote against fixing incorrect documentation; maybe my complaints gave them a change of heart, or maybe it was an honest mistake, but they did correct their vote. If you include a comment with qualitative information, at least Microsoft, the submitter, and the rest of the community have some insight into your decision.
In future installments, I'll try to stick to a theme and a certain number of Connect items (say, 5). What do you think? Suggestions?