Slashdot Mirror

← Back to Stories (view on slashdot.org)

Red Hat Reaping Benefits From Novell/MSFT deal?

Posted by Zonk on from the don't-like-their-dance-partner dept.

Ho Kooshy Fly writes "It seems that at least one software group has seen the use of Red Hat substantially increase over Novell as of late. EWeek theorizes that this may be backlash from the patent deal with Microsoft. From the article: 'The survey's findings can also be extrapolated to the broader open-source software industry and are not limited to those enterprise customers using Alfresco software "because of the wide range of open-source and proprietary software use cases captured and the large sample size of the survey," [Ian Howells, Alfresco's chief marketing officer] said. "We think these findings accurately reflect the broad technology trends across modern stacks in organizations of all sizes." Gallup polls about U.S. presidential candidates typically survey about 1,000 likely voters, while Alfresco surveyed more than 10,000 people, he said.'"

7 of 71 comments (clear)

  1. no bias? by vfrex · · Score: 4, Insightful

    I'm not convinced that this is unbiased, especially the claims that Novell adoption is slowing due to backlash. Give us the stats if you want, but let us draw our own conclusions. Also, why did the report separate Fedora and RHEL while not separating openSuSE and Novell SuSE? Certainly interesting data, but like I said, I don't really want their conclusions. On a side note, PostgreSQL seems to be used a lot more than I originally thought.

    1. Re:no bias? by BVis · · Score: 2, Insightful

      The patent deal I doubt would cause any impact on sales - I mean, unless there are high powered idiots making decisions based on emotive overtones rather than on a factual basis.
      You've just described every executive at every large company in the US. The executive washroom is a logic-free zone.

      If anything it'll be due to the same issue it has always had, terrible marketing, terrible sales people, terrible communication with customers and developers. The lack a leader with charisma and charm that can not only win over customers but developers as well. Everyone of them so far have either been technocrats trying to run a software company like a consumer product company or a geek who completely lacks any idea of business sense, marketing knowledge or the ability to sell products to non-technically minded people.
      That's its strength. It competes on its own, without the need for marketing bullshit. Think about it: this is a product that has made significant inroads on Microsoft's territory with no cohesive marketing message, no charismatic leader, only geeks trying to make good software for the good of the community. In the end the quality of the product is what should be important, not whose box is shinier.
      --
      Never underestimate the power of stupid people in large groups.
    2. Re:no bias? by kaiwai · · Score: 2, Insightful

      You're right. In theory, it should be nice for products to stand on their own merits, people look beyond the marketing hype, and customers do their own research rather than relying on biased studies funded by companies - be they Microsoft, IBM, Red Hat or Novell.

      Sun suffers from the same problem - awesome products marketing terribly; heck, its proven here with idiots making stupid statements that Solaris is only for SPARC, that Sun doesn't sell x86/x64 machines etc. etc. Ignorant goes both ways, and I'm always surprised to seeing the blatent ignorance that exists here.

    3. Re:no bias? by vfrex · · Score: 2, Insightful

      If that makes Linux a science project, I'm not sure what we should call an OS that doesn't bother attempting to play nice with others. I'm sure most sysadmin won't need to bother with dual booting Ubuntu next to Windows, somehow...

  2. A note about statistics by rar · · Score: 5, Insightful

    Gallup polls about U.S. presidential candidates typically survey about 1,000 likely voters, while Alfresco surveyed more than 10,000 people, he said.'"

    I recall that my statistics professor explicitly pointed out a common mistake in statistics: "Contrary to what people typically believe, the size of the sample is often not as critical as getting an unbiased sample." If you call the home phone number of people during daytime and ask the ones who answer whether they are employed or not, you will not get good statistics. Regardless if you reach 1 000 or 10 000 people.

    I have no idea if there is such a problem with the statistics presented here. I just want to point out their claim of sample size can not be taken to mean that their statistics are better than Gallup's.

    --
    Open Materials Database
  3. They remind of the british in india by badriram · · Score: 2, Insightful

    . It is a good stategy, and some open source people really do fall for it. GPL3 from a corporate perspective is not a very positive thing. So i would think that community is already showing lots of divide, although that could just be because there are lots more people in it now.

  4. Re:Statistics... by Phoenix00017 · · Score: 2, Insightful

    Indeed. I'm not sure why this point was made. n=1000 is generally enough to get statistically significant results. Having more than that in no way guarantees "more accuracy" or whatever they were trying to argue. What is important is the selection method. An evenly distributed, random, blind sampling of 1000 people is going to be far more accurate than a self-selected survey of 1,000,000 people with similar opinions. I'm not saying this is what happened in this case (or that Gallup polls are better), but ultimately the number itself is irrelevant to the accuracy of the report once you've crossed a certain threshold. You reduce your error, but you might simply end up with a small standard error centered on a biased response.