Announcement

Collapse
No announcement yet.

Purism Introduces Its Telepathy-Using GTK3-Based Phone Dialer Plans

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

  • #11
    Originally posted by grok View Post
    Regarding the Openmoko, well hardware sucked back then.
    Not really. Well, at least not for the reasons you provide. For 2006, when Freerunner was released, it was somewhere between middle and high tiers mobile hardware. There were just two exceptions - TI Calypso began to age pretty soon due to 3G rollout, and Glamo GPU turned out to actually be a decelerator when attached to "retina"-like high resolution screen that Freerunner had (years before iPhone), resulting in abysmal graphics performance.

    Originally posted by Hi-Angel View Post
    I vaguely remember reading that Telepathy framework is a mess to deal with.
    Also, its XMPP support is pretty poor these days. Jabber now is a pretty modern, mobile-friendly protocol, while Telepathy lacks any support of XEPs developed during last few years. Even if there was some activity to implement them in Gabble, it quickly turned out that implementing them in a way that won't be just a gabble-only, ugly hack, needs some changes to Telepathy's APIs... and that's where any work has stalled.
    dos1
    Phoronix Member
    Last edited by dos1; 18 May 2018, 04:31 PM.

    Comment


    • #12
      Originally posted by MagicMyth View Post

      What's so controversial about telepathy? Seemed like a good idea to me for all Free-desktop er desktop's to share a base communication framework. At one point we had Empathy and Kopete working nicely and you could run Google, Facebook, MSN, XMPP without issue and switch between desktops. Plus it allowed cool features like being able to establish an SSH connection via a contact without needing to configure port forwarding on a router. For some reason Empathy stopped being maintained and I think KDE forgot about Kopete (though I think it is still maintained).
      It *seemed* like a good idea, sure - but in practice, it was a nightmare. The ideas is simple enough, that you write an abstraction layer that encompasses all the different protocols, allowing one client to support them all more-or-less transparently. But the catch is that all of those protocols are different, and so your abstraction layer is either the lowest common denominator of features that be easily adapted, or it's the superset of every single protocol that you want to support, including all the contradictions between them.

      Consensus among the people who developed Telepathy is that it was a failed experiment - that building a layer that allowed clients to ignore the differences between underlying protocols was just not practical. In reality, the clients needed to know that stuff - needed to know if sending a file would go point to point or through a server, needed to know about whether a protocol supported named chatrooms or not, and whether they could be created on-demand. It just didn't work for more than the basics...

      Comment


      • #13
        Originally posted by Delgarde View Post

        It *seemed* like a good idea, sure - but in practice, it was a nightmare. The ideas is simple enough, that you write an abstraction layer that encompasses all the different protocols, allowing one client to support them all more-or-less transparently. But the catch is that all of those protocols are different, and so your abstraction layer is either the lowest common denominator of features that be easily adapted, or it's the superset of every single protocol that you want to support, including all the contradictions between them.

        Consensus among the people who developed Telepathy is that it was a failed experiment - that building a layer that allowed clients to ignore the differences between underlying protocols was just not practical. In reality, the clients needed to know that stuff - needed to know if sending a file would go point to point or through a server, needed to know about whether a protocol supported named chatrooms or not, and whether they could be created on-demand. It just didn't work for more than the basics...
        But then also worth mentioning, libpurple have had a better luck on the same matter.

        Comment


        • #14
          Originally posted by Hi-Angel View Post
          Thanks for find that Hi-Angle. Its a good (lengthy) read. This makes me think it would be great to have some kind of centralised experience place on tackling problems so we can all better learn as what McQueen wrote there is really good information and would likely help prevent other developers wasting time on attempted solutions that have all ready been tried. You can usually find a blog post or article here and there documenting a devs challenges but they are spread throughout the wide wild world of the Internet and search engines only get you so far.

          Let's hope the Purism guys learn from McQueen's experience.

          Comment


          • #15
            Originally posted by dos1 View Post
            As a baker, posts like this really don't give me much hope for Librem 5 Looks like they're closely following the mistakes of "a company who shall remain nameless" (Openmoko, if you don't get it). Such a simple proof-of-concepts should be a part of project's feasibility study, especially today, when we have projects like freesmartphone.org and oFono who already did the hardest parts. With a modem not chosen yet, looks like they're yet to discover the massive complexity of the stack that needs to be implemented there. What's more - Openmoko phones were perfectly capable of making phone calls when shipped. They just had a few rough corners that made them not exactly as reliable as one would expect. Its their journey that led to creation of things like FSO and oFono. The company should be named - while Openmoko itself failed, you're standing on the shoulders of its (and later Nokia's) work.
            At this point, I just hope for Librem 5 to be a good hardware for me to use with any OS I'd like to. Unfortunately, the past taught us that the default OS matters a lot, because if that won't be good enough, GNU/Linux phones will get yet another reputation hit and scramble to get enough interest to get rolling once again for years
            Great. Maybe you are a very experienced baker that makes delicious bread and other stuff, but this is about computer programming.

            What's your experence to criticize this? Are you a hobbyist but talented programmer like Con Kolivas?

            (joke)

            Comment


            • #16
              Originally posted by timofonic View Post

              Great. Maybe you are a very experienced baker that makes delicious bread and other stuff, but this is about computer programming.

              What's your experence to criticize this? Are you a hobbyist but talented programmer like Con Kolivas?

              (joke)
              Note that "baking" or "cooking" is the jargon used by some of the firmware modders, meaning "create a ROM based on binary blobs extracted by a bigger package or stock ROM", which is done by people not expert enough to recompile from source or in the situations where you just cannot recompile from source at all.

              Or at least it was some time ago, I stopped keeping track of that once CyanogenMod became a thing. Some examples from that period.
              https://www.xda-developers.com/dsixd...droid-devices/
              https://www.gadgetdaily.xyz/how-to-c...n-android-rom/
              https://forum.xda-developers.com/sho...d.php?t=885048

              Comment


              • #17
                Originally posted by starshipeleven View Post

                Note that "baking" or "cooking" is the jargon used by some of the firmware modders, meaning "create a ROM based on binary blobs extracted by a bigger package or stock ROM", which is done by people not expert enough to recompile from source or in the situations where you just cannot recompile from source at all.

                Or at least it was some time ago, I stopped keeping track of that once CyanogenMod became a thing. Some examples from that period.
                https://www.xda-developers.com/dsixd...droid-devices/
                https://www.gadgetdaily.xyz/how-to-c...n-android-rom/
                https://forum.xda-developers.com/sho...d.php?t=885048
                I knew it. Thanks, now you ruined my joke

                No problem, it's good to know people know this stuff

                Comment


                • #18
                  Originally posted by timofonic View Post

                  Great. Maybe you are a very experienced baker that makes delicious bread and other stuff, but this is about computer programming.

                  What's your experence to criticize this? Are you a hobbyist but talented programmer like Con Kolivas?

                  (joke)
                  As a matter of fact, I was maintaining one of the community distributions for Openmoko (and not only) phones for some time. It was OpenEmbedded based and used Bitbake's recipes to bake the rootfs images

                  I'm also not a native speaker. But yeah, something smelled fishy with this word, thanks

                  Comment

                  Working...
                  X