Cultural change at Microsoft, or, "What was the point in keeping quiet?"


In an anonymous comment on “Patch for tcserver.exe memory leak on Tablet PC – at last!”, zzz asks:



What was the point in “keeping quiet”? Is there some kind of NDA that stops from stating the obvious thing (holy cow we have a bug in our perfect product!) that everyone knows about? You could have instead tried to work with the tablet pc community – as it seems that it otherwise takes media attention to give such bug some deserved attention.


This is a good question, and my reasoning probably deserves a wider audience than just a comment-reply, so here goes. There’ll be a plea for feedback at the end 🙂



  1. The Tablet PC community actually knew about the bug and the kill-two-processes-then-restart-one workaround back in May 2004 – see this link from Orlowski’s original article. So I wasn’t saying anything new, I was just saying it as a Microsoft employee. Despite/because of this, my post rapidly became the top Google hit for “tcserver.exe”.

  2. No, there’s no NDA, but there does seem to be a general unwritten rule that you don’t pile on to bug reports about another team’s product. If it had been my own team, or the Tablet PC team had been blogging back then, I’d have been much more aggressive. Instead I decided to work from the inside, rather than the outside.

  3. If you break rule #2, you run the risk of getting unwelcome notice from the other team, journalists (and then our public relations people), and finally your boss. Sometimes this is worth it, sometimes it isn’t, but ultimately it’s a personal judgement call. I’ve skirted close a couple of times over the past year and figured my boss was losing patience with giving me political air cover, so I let this one slide…

  4. Pre-announcing patches is almost always a bad idea, because they typically take longer to test than you first think. This definitely happened here. Do you think the anonymous spokesman who said this back in February is feeling good about himself?

    “Progress is being made and there’ll be an update soon,” a spokesman told us. “For now, we advise users to reboot the machine on a daily basis,” he added.



  5. Finally, media attention does work as a forcing function in these situations. Media attention gets executive attention, which in turn percolates down and results in change priorities. It’s a fact of life.


Here endeth the self-justification. Am I happy about how long it took to get this fixed? No. Would I do it the same way again? No – but that’s because we have at least one Tablet PC tester blogging, and I could engage on his blog rather than going through “the usual channels”. Do I think I did the right thing at the time? Probably — at least if I wanted to keep blogging 🙂


What do my readers think? Should I have risked all back then? Pushed the Tablet PC team for greater transparency? Done so from the inside or the outside? In short, how do you make cultural change happen in your job?

Comments (9)

  1. You’re quite right, that we suffered from 4) and that 5) does work as a forcing function. We wouldn’t have minded 2) as it might have helped our case for fixing the issue.

    I work closely with the remnants of the Tablet SE team (both testers joined us on the Platform test team) and the process of getting QFEs evaluated by Dev, Test, Management; then approved by Executive Management, war, etc is a long one.

    I am happy to hear anytime from any Tablet customers, internal or external, because if you have a comment, it means you’re using my product. If you have no comments, either my product is perfect, or it’s not being used. <big grin>

    I’m relatively new to blogging, and have been cautioned to watch what I write because of the sheer volume of NDA and unannounced "stuff" that I work with. That will change when Longhorn becomes more readily available.

  2. Thanks Hilton – yeah, I’d forgotten to add: "#6. The Tablet engineers I worked with all really wanted to fix this bug" 🙂

    Interesting to hear that you wouldn’t have minded more pressure early on. Maybe the right thing to do in the future is to sound out team members privately, to see if more "outside" pressure would help things…

  3. http://www.eweek.com/article2/0,1895,1834501,00.asp

    "…Microsoft calls the patch a hotfix that may receive additional testing and recommends that users who are not severely impacted by this issue wait for the upcoming Windows XP Service Pack due out on August 9 rather than download the patch at this time."

    "…informing its customers that they should wait to apply it unless they are having a problem with the leak."

    —————————

    Not severely impacted? Everyone is impacted, who is NOT having this problem? They release a fix, and now are saying to wait? May receive additional testing? Must mean they KNOW it’s buggy? For the life of me I can’t make sense outta this. Windows XP Service Pack? Can’t be XP SP3…so 2005 SP1? What on earth? Well, Media Center is getting an update bundle this fall, so maybe there IS some sort of Service Pack, Tablet PC + Media Center hooks/combo?

    Here’s a fix, for a nearly year long problem…oh, wait, don’t use it, wait until some yet-to-be-announced Service Pack? Man…maybe I don’t have all info, but this sure looks like a major trainwreck.

  4. Christopher – I think someone at eWeek is confusing "Service Pack" with "monthly patch update". August 9th is the second Tuesday of the month, i.e. the date when we release scheduled patches.

  5. Ok that makes sense (I think)…wait for the automatic "monthly patch update" download. We fixed it, but don’t rush out and fix anything just yet, wait for the full-package-with-all-other-fixes? With the subtext of ‘we should have waited until then anyways’? Ok, understandable and reasonable.

    But, the "unless they are having a problem" implies that some aren’t. And indeed, heard from a number of internals that said ‘they’ have never experienced it and that the press is simply making too much of it. Letsee the same code in *all* Tablets, impacts some but not others? How can that be? So in effect, saying it’s OEM related? Surely not. I am confused. 🙂 Maybe just damage control.

  6. Chris,

    Please don’t try to make more of this QFE than there is. "Reading between the lines" may be your vocation/passion, but in an organization the size of Microsoft, there are many different hands that every piece of official communication and every bit of code must pass through before being made public.

    In my view, each and every point you’ve tried to make in this thread is a variation on "what are you covering up?" and the answer is "Nothing". Once the code leaves the product team’s hands, it is vetted and cleaned up by other hands, it’s out of our control.

    I truly do want each and every Tablet user to install this patch, whether from downloads.microsoft.com or windowsupdate.microsoft.com or wherever. It will make their future Tablet usage better. On the other hand, many of our users do not experience any problems from the leak in the first place, so why should they patch?

  7. Ahh yes, I am way too overanalytical sometimes, I will grant you that. Just been a long frustrating process, though not a super big deal in the grand scheme of things. It’s just that my patience ran out at about the halfway point, and that was when nothing was being said. I guess I got lucky, as this problem was very prevalent to me, uptime styled and using TIP lots, I guess. But in the final analysis, glad it’s here, as you did listen, even tho way too long coming. =) And it hasn’t lessened my feelings towards the Tablet in any form, still the biggest fan. The Tablet concept will survive any glitches.

  8. zzz says:

    How does "each and every" become "why should they patch?". Contradiction.

  9. Orlowski’s noticed: "How the US military chivvied up Microsoft" (http://www.theregister.co.uk/2005/07/13/tablet_pc/).

    And although Orlowski’s story appears to be total speculation based on Peter’s original rantlet, I for one am all in favor of regular fly pasts of the Redmond campus 🙂