And the other problem with Linux manifests itself
by James on Feb.05, 2010, under Computers
Woke up this morning and this was sitting in my Twitter feed.
@PickettBill: RT @jforestier http://ljforestier.com/aB Not ready? Ha, is perfect for # of users. Dump the # of Windows users on Linux things fixed fast.
This is why the usability problems will not be fixed, “I’m sorry, it’s not me, it’s you.”
I think it’s time to stage an intervention.
Tweet this!
Related (maybe) posts:
February 5th, 2010 on 17:10
Seriously dude, ;) Tweet back. Much easier to keep track of than collating feeds to come across your blog! =D
February 5th, 2010 on 17:13
Don’t see that post in your stream or my mentions. Meh, maybe twitter is twitting here! ;)
February 5th, 2010 on 17:25
Ok, here is what I meant not what is confined to 140 characters to interpret in Twitter ;) Linux is doing quite a bit of what it does without cooperation from other parties. The Open philosophy and process is much more democratic and plural than anything Closed philosophies have to offer. Even if at this point the actual offerings may be of differing qualities. Microsoft does not need to write their own device drivers for everything someone decides to put out for Windows, the loose confederation that is “Linux” often does have to. The field of development is huge in Linux, not everything gets the love it deserves at every moment. Considering that only a very small fraction of people contribute to “Linux” which includes everything, not just the kernel (actual Linux) but everything in Userland too like the Gnome desktop environment then it is no surprise that making it go takes more than has been committed in areas. When Linux has the sheer critical mass of users that Windows currently enjoys these contributions will fall into place: at the very least there will be better guides to find the errors simply because more people will be reporting them. You are absolutely correct in the respect you have focused on: right now, usability and completeness could be improved for Linux. The issue I’d like you to address is that regardless of what exists in this moment, Open is a better development process than Closed.
February 5th, 2010 on 22:03
I agree that an open development process is far superior, but I also feel that it must be, in some ways, centralized.
Why does there need to be rpms AND debs? Why not one or the other (or even a completely neutral third option).
Development environments should be standardized and attempts at backward compatibility must be made.
For example, the recent problem I ran into with FUPPES was that after tracking down all of the various dependencies, the source code was not able to be compiled with my version of GCC as it was not compatible with the older version that the developer uses. Of course, pre-compiled binaries would fix the whole problem.
Personally I like the compromise that Google has come to with Android and, presumably, Chrome OS. An open development system with a centralized authority that allows for one development environment and a standard application installation.
To a certain extent, the distros that are available attempt to do this but because there is no enforcement mechanism, devs are free to ignore them as they see fit.
There is a difference between openess and anarchy.
February 5th, 2010 on 23:10
I *have* experienced the pain as well. It is some comfort to compare the whole Linux F/OSS scene to an evolutionary system. In The Cathedral and the Bazaar Eric Raymond gives a good overview. The price of giving up centralization is an anarchy of sorts but by the same token the freedom to explore beyond what a small group thinks is “the” way is strength through diversity in the long-term. Instead of wholly embracing either of these extremes I opt that a middle-ground should be found. The Linux Standards Base represents such a position. Instead of each distribution being forced to re-invent the wheel on every birth, view it as a cycle of explore and compact. The proven systems get compacted into the LSB while every distribution is still free to perform “bazaar” functions and grow the entire ecosystem in unexpected ways. The “cathedral” benefits mostly in the short-term. In it’s time-scales amazing systems can be imagined, shaped, and brought to fruition on the side of a dime. It is seductive, it gives the feeling that “progress” is being made. It is also a trap. The cathedral is slow death, the democratic and most important *plurality* of the bazaar is continual adaptation. The weakness of the bazaar is the time-scales involved. We agree on a point, more centralization of a kind is needed in Linux-land. My preference for how it should emerge is that it should be optional and motivated by practicality. The LSB exists, you are not forced to follow them. If LSB turns out to be an evolutionary dead-end well at least some of the distributions didn’t go that route. The cathedral is not automatically doomed either to a dead-end but once the decision has been made at the top it is a lot easier for the cathedral to unwittingly commit to that unfortunate extinction. In the mean-time I wish I understood why my USB hard drive unmounts about every 24 hrs ;) Ah well, its not perfect but I love it anyway =D
http://catb.org/esr/writings/homesteading/cathedral-bazaar/
http://www.linuxfoundation.org/collaborate/workgroups/lsb