Announcement

Collapse
No announcement yet.

What Would You Like To See Next From Phoronix?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #41
    Originally posted by cl333r View Post
    1) .deb installer for Ubuntu of latest phoronix benchmarking test suite from site.
    I would suggest you take a look at http://open-build-service.org/ it can build for a lot of systems and highly automated, so you can offer rpm or deb or whatever else there is

    Comment


    • #42
      a few thoughts, maybe I'll have more later on

      * Less fixation on Ubuntu.
      * better login support for Premium Membership
      * Michael gets VIA kicked to release specs and a decent free driver (I can still dream)
      * all games will also be published in a Linux native version (c'mon!)
      * not too much layout changes. Maybe even though there might be possible improvement here or there I got used to it and I kinda like it as it is now
      * maybe a few more Hardware tests? And not just in terms of pure benchmarking, more like running at all (though there'd be the question if that collides with e.g. tuxmobile etc.)
      * less speculation, more facts


      * Michael starts to drink beer from the region of Saxony (.de, adjacent to Bavaria) and Bohemia (.cz, same for this) too.
      (the Bermuda triangle of beer)
      Stop TCPA, stupid software patents and corrupt politicians!

      Comment


      • #43
        benchmarks for OS performance

        Most of the Phoronix benchmarks compare the efficiency of application code and compilers, i.e. newer releases compared to older releases. The OS performance is not being measured. By OS "performance" I mean the effectiveness of the OS in getting the best overall performance when running competing processes contending for resources, not when running one application for a benchmark of that application. I often notice that an interactive process (GUI or keyboard) that uses very little resources will have poor responsiveness when other (batch) processes have saturated the CPU or disk bandwidth. A good OS scheduler would boost the priority of a low-resource process (basically a human-bound process using little CPU or disk capacity). The ability of an OS to remain responsive under load is one measure of OS quality. I would like to see some measure of this for the Linux kernel and others (BSD, Apple, Windows, Andriod). This is no trivial task. A "standard load mix" of contending processes would have to be developed along with a standard interactive process to measure. The benchmark would measure the overall throughput of the standard load mix as well as the responsiveness of the interactive process.

        Comment


        • #44
          I am a Phoronix fan. I like the speculation (to a certain degree), the mailing list summaries and the occasional bleeding-edge "inside" info.

          What I would like to see next is the following:

          - More writers, as mentioned many times before. Perhaps you can allow user-submitted articles (if they pass your quality control)?

          - More extensive conclusions in the benchmark articles. This will make the benchmark itself more interesting. Many of the benchmark articles just show some (more or less) nice graphs with a brief comment to each of them and a very short summary at the end. If there is nothing more to say than a brief paragraph to four pages of graphs, then perhaps it is not necessary to post all the details?

          - Proper login for premium members. This was quite an annoyance when I was a premium member. I always had to go from an article, to the forums and then back to the article to benefit from the premium membership. I would love to be a premium member, but currently it's too annoying.

          - Lose the kontera ads. They are the worst kind of ads, popping up in front of the text I'm trying to read. These ads on Phoronix are the sole reason for why I had to install adblock (I only block the kontera/intellitxt ads though). I would prefer to be a premium member instead, but as mentioned above it's too cumbersome.

          Comment


          • #45
            Originally posted by Xilanaz View Post
            I would suggest you take a look at http://open-build-service.org/ it can build for a lot of systems and highly automated, so you can offer rpm or deb or whatever else there is
            Ya now there is a service worthy a 5 or 6 page article in itself.

            Comment


            • #46
              Originally posted by Michael View Post
              Me... For news, featured articles, and everything else.
              May be keeping things small is a good idea. I hate big numbers too
              Great job, anyway!

              Comment


              • #47
                a few more ideas

                +1 for flattr button and bitcoins cause i'm not much into paypal and i i'd like as individual to offer you a ''beer'' once in a while for your excellent job.
                +1 for a phoronix spin off on wireless cards!!
                +1 for getting some donations and sponsors and hire people to help cause burn out isnt pretty...

                Comment


                • #48
                  Full-Length RSS Feeds for Premium Subscribers

                  I monitor Phoronix thru Google Reader, and it is annoying to have to click thru to read the entire article. Maybe specially for premium subscribers Phoronix could include the entire article in the RSS feed. I think Ars Technica offers this feature to it's paid subscribers.

                  Comment


                  • #49
                    1) Fix the login problem with article pages. Once per day, I do this: I load an article, see it's page 1 of 4, open forums in another tab, wait for it to load, then close it without even looking at it, and then reload the article. Aah, page 1 of 1, that's better. I just wasted my time, and wasted your computer's time too. Poor computer. Do it for him, if not for us.

                    2) Include oldschool filesystems in filesystem benchmarks. Yes, I know ext4, btrfs, and zfs are where all the fun is, but showing the numbers next to Reiser3, xfs, and even ext3, and especially the constantly-underrated jfs, helps put things in perspective. If one of the new ones start beating reiser3 at lots-of-small-files, or even gets to be half as fast, I want to know that my next mail server deserves some new thought.

                    Comment

                    Working...
                    X