Is Linux Maintainable?
 

No trolling intended. Just looking for honest comments about this lecture.
https://youtu.be/eXcFA_JvBOg
Thanks.

Honest comment...

I opened the URL and when I saw that it was a Micro$oft video I knew that I had already wasted that much bandwidth and closed the browser tab to limit my losses.

Watching it would be a complete waste of time and bandwidth.

"Longitudinal study of 365 versions of linux " was enough for me.

Yes, Linux is maintainable.

Are there any other points, questions related, examples, or citations needed?

Is Linux Maintainable?
 

If we just consider the variables defined on the study, of course his case looks compelling.
However, I suspect that he is missing other relevant factors. For instance, the developers must have good reasons to let Cases 4 and 5 (especially 5) crop up so frequently. And particularly, I'm sure there must be significant advantages from raising the density of definitions per line on the kernel as described on the lecture.
Finally, how come Gnu/Linux is such a success in some areas of business, academia and government despite what he portrays as fundamental flaws.

Member response
 

Hi,

I watched about 5 minutes of the video and realized that the modeling was slanted towards a Microsoft model of maintenance. Therefore the adherence or adoption of Microsoft's maintenance would not fit the open-source model.

Of course the involvement of participants can be restricted by the maintainer but that too can be circumvented by someone opening a parallel project.

So my answer to your query is; Sure Linux is maintainable and is currently done by multiple maintainers within the environment with project management via the individual projects that support Linux. So to apply this to Gnu/Linux then the maintenance is done within each project that accums to be be a distribution thus someone's OS.

I really did not wish to waste anymore of my time viewing your linked video. Microsoft really does not like the Open-source community and will do anything to bring the community down.

By not watching the entire video then take my comments as my views only as related to Microsoft's past attempts for wishing Linux or Gnu/Linux ill will.

Hope this helps.
Have fun & enjoy!
:hattip:

Thanks, onebuck.

I kept that in mind while looking at the presentation.

Anyway, since this particular presenter is better spoken and more civil while referring to the Free Software community, I decided to judge his message by its merits, from my perspective as a regular user who is, however, quite conversant with general research methodology.

I have to admit that my original post was driven by more than simple curiosity.
Actually, although my personal experience as a Linux user for years, has been very positive (see my signature below), my experience with system administrators for my business has been consistently disappointing (I wonder if other small non-IT-business owners have different experiences); of course, that is a subject for another post I will make in the future.
Long story short is that, instead of completely blaming it on their ineptitude and/or unprofessionalism, I was hoping to give them the benefit of the doubt and consider that maybe the cause was the inherent difficulty to administer Linux; that was why I was hoping that lecture to bear some weight, in order to (in my very personal experience) absolve these deplorable individuals I've had the misfortune to meet.
Again, no offense to anybody, just my 2 cents.

365 versions?? kernels?? distros??

I maintain development, office, ASP operation, and web interface machines based upon Linux (mostly RHEL and CentOS). I have since RHEL4 was new, and we are now mid migration to RHEL7. Maintenance of these systems has been a breeze, far easier and much cheaper than maintaining the Windows servers. The Microsoft side of the house requires more downtime, more man hours, and is more difficult and costly to maintain.

In another environment we ran all Debian, because we like the philosophy, stability, and specific package options. It was the same in this respect: it was far cheaper and easier to maintain than Windows, and with far less downtime.

Before Linux we ran SYS-V (AT&T sysv version 3 on 3b2 and 3b15 machines) and HP-UX. Back then you had to add Novell (With Novell TCPIP packages) to Windows Server to get any network integration. Again, Unix was cheaper and easier to maintain than Windows (though Novell was OK and helped a bit).

My experiences over the last 34 years may be atypical, but I suspect not.

Its simple.
MS made a software product that made it a LOT of money.
so when a free (cost-wise), relatively better operating system is available , it hurts MS business.
MS always had a intention of killing its competition, but Linux never had that intention, it just evolved leaps and bounds out of a hobby project.

absolutely and why shouldnt it. MS is a huge empire which has many mouths to feed. One way is to spread FUD among users.

While I have no enterprise admin experience to speak of I know at home I used to spend almost an hour per computer in my house on Windows before I switched. Updates, scans, defrags... the works. Of course updates required a few reboots. Now I spend 10-15 minutes total between my server and 2 laptops. Updates are all that is required and reboots are few and far between. I don't do much other than that for actual required maintenance.

Linux is maintainable for as long as people maintain it. The fact that it's written in a programming language which doesn't discourage bugs and undefined behavior, or even curb global mutable variables doesn't seem to have slowed it down much.

...


I found the lecture that I watched to be very academic.

It sets up criteria to judge software in a very superficial sense without any context: count the global variable 'current' and analyze how it gets used (mutated or not). You could do that about any part of the kernel/C: count occurrences of free vs malloc; lines of code; if statements without braces; gotos; &c, and we probably wouldn't learn how maintainable anything is. It's difficult to take their word for it that 'global coupling' is A Very Bad Thing without any supporting evidence. Maybe things aren't as simple as they would like it to seem.

The lecturer admits early on that in this context "maintainable" can't be defined, but then later on flat out lies and says that unless some supposed bad construct is dealt with, Linux will not be maintainable. Maybe that was just a slip, but what can we expect if from the start we're disallowed from any meaningful conclusion?

I wish I hadn't watched it, because it's a little less exciting than watching paint dry, and there's not much to learn from it other than some people spent a lot of time making a study and writing a paper about something superficial, back in 2005.

Hey, let's keep an open mind about this . . . !! :D

For instance, what about Red Hat's business model? You pay them to provide updates for you, and "Scout's Honor" that you won't monkey around with the source-code yourself even though you can (mostly) have it. With the collective revenue thereby obtained, RHAT (a very healthy publicly-traded company ...) pays a large number of developers to fulfill that promise. The business-model works because their efforts are amortized over a very large number of subscribers.

Similarly, Ubuntu. Thanks to the generosity(?) of Mark Shuttleworth, they furnish very-current updates to you at no charge. But, they also have a for-subscription service, similar in many ways to Red Hat's, which IMHO is ample indication that Mark is a shrewd businessman, not just a generous man. (Which, I hasten to say, he also is. Thank you, Mark.)

IBM? Heavily involved in Open Source even though they own several proprietary systems. Everything Microsoft ever learned about the software business, they learned from IBM.

How about MySQL? Did you know that they have a proprietary version of their product, also?

Bottom line (IMHO): "Yes, Linux is 'maintainable.' Yes, Linux is 'no accident.' Yes, you can bet your business on it." But the "by subscription" side of it, again IMHO, is not mutually-exclusive with the "free as in beer." The two aspects of Open Source are critically joined at the hip.

This is a business model that "never before has existed, because never before could it have existed." But, it works. It has dramatically changed the once-prohibitive cost structure of software development and it has thereby enabled software innovations ... and a pace of software innovation ... that never before could have existed. Everyone is reaping the benefits, even the "proprietary" companies.

Also, IMHO, it's a business model that quite-naturally has "multiple arms." It shows itself capable of supporting both the needs of (sometimes, gigantic ...) businesses, and the needs, even, of hobbyists. It has made software available ... free(!) of charge ... that once could not have been obtained at any price.(Because no one thought that they could afford to build it. And they couldn't have ... on their own...)

Hi.
 

For sure and free; we all need or are the admin, never trust that a operating system or other software will do this for you! ;)

Have fun! :hattip:

Attacking Linux because it's complex? What about the firmware for defense systems, aircraft, spacecraft? What about financial software used to track Wall Street trading?

Sit there and attack an operating system because it has a lot of files and variables. Nice! How many files does Windows have as part of it's source repository?

Oh right! We don't know and they're not going to reveal that!