4 ms·

This is gold: http://www.reddit.com/r/IAmA/comments/dkk3l/iama_we_are_members_of_the_ie9_product_team_here/c10vmml http://www.reddit.com/r/IAmA/comments/dkk3l/i

This is gold: http://www.reddit.com/r/IAmA/comments/dkk3l/iama_we_are_members_of_the_ie9_product_team_here/c10vmml http://www.reddit.com/r/IAmA/comments/dkk3l/iama_we_are_memb... The MS guy admits that hardware acceleration can be turned off in the config menu after saying IE9 can't run on windows XP because it requires hardware acceleration.

pavs 16y ago
At this point I am starting to feel sorry for the MSFT guys. Its quite obvious that the engineers don't run the show, they just do whatever the profit-making overlords wants them to do.
- patrickaljord 16y ago
  I don't feel sorry for them, they are talented engineers, that means they could work for any company on any project they want. They chose to work on IE for MS.
- djacobs 16y ago
  I agree, the MSFT engineers (not the marketers) seem likeable enough. I think they're handling criticism reasonably enough and I think the Reddit crowd should understand exactly how little input these engineers get inside of Microsoft's development model. (Still, I will continue recommending Chrome and Firefox to friends. IE anything is just bad.)
  - sriramk 16y ago
    Disclaimer as usual: I dont speak for MSFT, yada, yada. I rarely get into one of these 'defend-MSFT' threads but I couldn't let one more of these 'MSFT engineers are great but marketing is not' threads go by. I think your characterization of how product managers and devs/PMs/test work at MSFT is way wrong and represents a simplistic 'engineers vs bean counters/marketing' of sorts (apologies if that's a strawman). It's always a tag team effort and it's not always easy to see the boundaries between marketing/biz dev/engineering. All sides take part in collective decisions. If you have any real issues with big product decisions for any product at Microsoft, you need to go to a VP or the executive team. And most Microsoft VPs are from a engineering lineage. On a separate topic, I have nothing to do with IE but in this case, the product managers seem to be working really, really hard with what is a extremely hostile crowd. In some cases, the community has assumed what the real answer should be and even when they get the absolute straight line on why something the way it is, there is disbelief. (the truth as what I would hear if I met Dean Hachamovitch in an elevator and asked him the same question as a Microsoftie).
    - pavs 16y ago
      To be fair about the criticisms, it seemed to me like somebody made a list of things they could possibly do wrong while doing a reddit AMA (in front of a bunch of hardcore tech people) and then went on making those mistakes. What were they expecting when they decided to do an AMA on Reddit? If their main goal was to make more people try IE9, and stick with it, I think they did a really horrible job.
      - djacobs 16y ago
        I agree, I think the marketers failed at their own game (understanding their audience and selling themselves) today/yesterday. They sidestepped questions like politicians. If they had understood their audience, they'd have known this would not work, that details and direct answers matter. As it is, they just revealed Microsoft's "feature" model—i.e., we only implement technology that is popular now ... on websites that are built around the lowest common denominator of web technology. IE6. I stand by my statement that the programmers seem well-intentioned and honest (direct, even), where the marketers are clearly avoiding tough questions.
        sriramk 16y ago
        I have done several customer engagements in past teams from conferences to very private executive engagements where I've been the 'technical guy'. Since most questions flying at me are factual, it is easy to respond with factual responses. But with any hard question which goes outside pure tech, I would deflect it some poor biz-dev/product manager who would have to weigh several things in his or her response - Microsoft's strategy, future plans (both disclosed and secret), competitors and their selling points, what this current customer expects to hear among a host of other things. It is a much harder job and I often came out looking like the 'technical guy who was a straight shooter, not like that other suit' though I had the much easier task. This isn't some Microsoft thing alone. If you parse through any large scale engagement from any big company, you'd almost always see this pattern. Go to any conference and walk up to a large company's booth. If you ask technical questions, you'll often get the engineer to respond. If you ask some roadmap/competitive positioning/touchy question, you'll get directed to someone else. This is usually because the engineer involved doesn't know the answer/strategy/market positioning/whatever. To say that one group is somehow more well-intentioned/honest than the other seems wrong.
jfager 16y ago
To be fair, they've been saying IE9 requires D2D, which provides both hardware and software rendering. All he's saying here is that you can ask IE9 to use D2D's software rendering if your VM is having problems with hardware rendering. This is not the same as admitting they don't really need D2D.
- patrickaljord 16y ago
  > This is not the same as admitting they don't really need D2D. It is admitting they don't need hardware acceleration, which is what they advertise.
  - jfager 16y ago
    Given that we've had web browsers since 1992, and that they've only recently started supporting hardware acceleration, that doesn't seem like something that needs to be "admitted". They advertise it because when and if it works for you, it is a feature. I'm not pro-MS here, I promise. But they suck hard enough on real issues that it's stupid to go after them for pretend ones. If you're going to slag them on this, slag them for deciding to depend on a Vista+ api. Don't pretend they're accidentally revealing their own lies when they aren't.
  - djacobs 16y ago
    You're right, it's what they advertise officially, but now that we're talking to engineers, we get the gritty details, a story I accept: D2D is a layer that doesn't exist in Windows XP. Where it does exist (Vista and 7), you can use it to switch between hardware- and software-accelerated rendering. Where it does not exist, they would have to rewrite entire libraries or APIs to interact with the XP kernel. According to the answerer, rewriting these libraries would amount to rewriting IE9 as IE8 with a GUI upgrade. (That the graphics rendering sits inside the kernel is another problem, but not likely one these guys can fix.)
    - noibl 16y ago
      > According to the answerer, rewriting these libraries would amount to rewriting IE9 as IE8 with a GUI upgrade. I couldn't find this in the discussion. It also seems unlikely given how tangential D2D is to most of the interesting features in IE9.
jbeda 16y ago
Having worked on IE back in the day, I can take a good guess as to why XP isn't supported. (I haven't had the time to read all of the reddit responses yet). It comes down to cost. Microsoft tests everything way way more than is apparent from the outside. Adding XP support (backporting some version of software D2D) is probably not that difficult from an engineering perspective, but the ongoing test cost is significant. That functionality/cost tradeoff is largely out of the engineering teams hands. The chain of 'whys' probably goes like this: Q: Why can't they do XP? A: Because D2D isn't on XP and because of test cost. Q: Why is test so expensive? A: They have to test the hell out of it and XP increases the matrix. Q: Why so much testing? A: Because of long release cycles means you only have one chance to get it right. Q: Why long release cycles? A: IE is part of Windows and OEMs and Enterprises want long lead times to qualify software. Q: Why is IE part of windows and not disjoint like all other browsers? A: IE has to roll up to some P&L in Microsoft and the mission is to make Windows a more compelling platform. BTW -- Safari 5 doesn't run on Mac OS X 10.4...

pavs 16y ago

At this point I am starting to feel sorry for the MSFT guys. Its quite obvious that the engineers don't run the show, they just do whatever the profit-making overlords wants them to do.

patrickaljord 16y ago
I don't feel sorry for them, they are talented engineers, that means they could work for any company on any project they want. They chose to work on IE for MS.
djacobs 16y ago
I agree, the MSFT engineers (not the marketers) seem likeable enough. I think they're handling criticism reasonably enough and I think the Reddit crowd should understand exactly how little input these engineers get inside of Microsoft's development model. (Still, I will continue recommending Chrome and Firefox to friends. IE anything is just bad.)
- sriramk 16y ago
  Disclaimer as usual: I dont speak for MSFT, yada, yada. I rarely get into one of these 'defend-MSFT' threads but I couldn't let one more of these 'MSFT engineers are great but marketing is not' threads go by. I think your characterization of how product managers and devs/PMs/test work at MSFT is way wrong and represents a simplistic 'engineers vs bean counters/marketing' of sorts (apologies if that's a strawman). It's always a tag team effort and it's not always easy to see the boundaries between marketing/biz dev/engineering. All sides take part in collective decisions. If you have any real issues with big product decisions for any product at Microsoft, you need to go to a VP or the executive team. And most Microsoft VPs are from a engineering lineage. On a separate topic, I have nothing to do with IE but in this case, the product managers seem to be working really, really hard with what is a extremely hostile crowd. In some cases, the community has assumed what the real answer should be and even when they get the absolute straight line on why something the way it is, there is disbelief. (the truth as what I would hear if I met Dean Hachamovitch in an elevator and asked him the same question as a Microsoftie).
  - pavs 16y ago
    To be fair about the criticisms, it seemed to me like somebody made a list of things they could possibly do wrong while doing a reddit AMA (in front of a bunch of hardcore tech people) and then went on making those mistakes. What were they expecting when they decided to do an AMA on Reddit? If their main goal was to make more people try IE9, and stick with it, I think they did a really horrible job.
    - djacobs 16y ago
      I agree, I think the marketers failed at their own game (understanding their audience and selling themselves) today/yesterday. They sidestepped questions like politicians. If they had understood their audience, they'd have known this would not work, that details and direct answers matter. As it is, they just revealed Microsoft's "feature" model—i.e., we only implement technology that is popular now ... on websites that are built around the lowest common denominator of web technology. IE6. I stand by my statement that the programmers seem well-intentioned and honest (direct, even), where the marketers are clearly avoiding tough questions.
      - sriramk 16y ago
        I have done several customer engagements in past teams from conferences to very private executive engagements where I've been the 'technical guy'. Since most questions flying at me are factual, it is easy to respond with factual responses. But with any hard question which goes outside pure tech, I would deflect it some poor biz-dev/product manager who would have to weigh several things in his or her response - Microsoft's strategy, future plans (both disclosed and secret), competitors and their selling points, what this current customer expects to hear among a host of other things. It is a much harder job and I often came out looking like the 'technical guy who was a straight shooter, not like that other suit' though I had the much easier task. This isn't some Microsoft thing alone. If you parse through any large scale engagement from any big company, you'd almost always see this pattern. Go to any conference and walk up to a large company's booth. If you ask technical questions, you'll often get the engineer to respond. If you ask some roadmap/competitive positioning/touchy question, you'll get directed to someone else. This is usually because the engineer involved doesn't know the answer/strategy/market positioning/whatever. To say that one group is somehow more well-intentioned/honest than the other seems wrong.

jfager 16y ago

To be fair, they've been saying IE9 requires D2D, which provides both hardware and software rendering. All he's saying here is that you can ask IE9 to use D2D's software rendering if your VM is having problems with hardware rendering. This is not the same as admitting they don't really need D2D.

patrickaljord 16y ago
> This is not the same as admitting they don't really need D2D. It is admitting they don't need hardware acceleration, which is what they advertise.
- jfager 16y ago
  Given that we've had web browsers since 1992, and that they've only recently started supporting hardware acceleration, that doesn't seem like something that needs to be "admitted". They advertise it because when and if it works for you, it is a feature. I'm not pro-MS here, I promise. But they suck hard enough on real issues that it's stupid to go after them for pretend ones. If you're going to slag them on this, slag them for deciding to depend on a Vista+ api. Don't pretend they're accidentally revealing their own lies when they aren't.
- djacobs 16y ago
  You're right, it's what they advertise officially, but now that we're talking to engineers, we get the gritty details, a story I accept: D2D is a layer that doesn't exist in Windows XP. Where it does exist (Vista and 7), you can use it to switch between hardware- and software-accelerated rendering. Where it does not exist, they would have to rewrite entire libraries or APIs to interact with the XP kernel. According to the answerer, rewriting these libraries would amount to rewriting IE9 as IE8 with a GUI upgrade. (That the graphics rendering sits inside the kernel is another problem, but not likely one these guys can fix.)
  - noibl 16y ago
    > According to the answerer, rewriting these libraries would amount to rewriting IE9 as IE8 with a GUI upgrade. I couldn't find this in the discussion. It also seems unlikely given how tangential D2D is to most of the interesting features in IE9.

jbeda 16y ago

Having worked on IE back in the day, I can take a good guess as to why XP isn't supported. (I haven't had the time to read all of the reddit responses yet). It comes down to cost. Microsoft tests everything way way more than is apparent from the outside. Adding XP support (backporting some version of software D2D) is probably not that difficult from an engineering perspective, but the ongoing test cost is significant. That functionality/cost tradeoff is largely out of the engineering teams hands. The chain of 'whys' probably goes like this: Q: Why can't they do XP? A: Because D2D isn't on XP and because of test cost. Q: Why is test so expensive? A: They have to test the hell out of it and XP increases the matrix. Q: Why so much testing? A: Because of long release cycles means you only have one chance to get it right. Q: Why long release cycles? A: IE is part of Windows and OEMs and Enterprises want long lead times to qualify software. Q: Why is IE part of windows and not disjoint like all other browsers? A: IE has to roll up to some P&L in Microsoft and the mission is to make Windows a more compelling platform. BTW -- Safari 5 doesn't run on Mac OS X 10.4...