[gentoo-dev] Re: College Course in Gentoo Development

[Duncan]

Paweł Hajdan, Jr. posted on Mon, 17 Dec 2012 08:48:01 -0800 as excerpted:

> On 12/17/12 7:32 AM, Anthony G. Basile wrote:
>> So what should I teach?  Here's what I've got off the top of my head:
>> Please comment.  If it gets systematized enough, it can be a guide to
>> future devs too.  Everything will be creative commons.
>
> I think it's worth to mention somewhere that although packages take
> longer to compile than downloading binaries, people don't have to
> _watch_ the compilation, and many things can be done e.g. overnight.

++

Two reasons.  From a user perspective, I often see people's complaints 
about how much of a waste of time gentoo is, watching the build output 
scroll by.  People simply don't seem to realize that you can go away and 
do something else while all that's happening, or on a modern multi-core 
system with a decent amount of memory, set PORTAGE_NICENESS and MAKEOPTS 
appropriately, and even just go on using the system as you normally would 
if you weren't compiling in the background (of course as long as that 
"normally" isn't something like multi-thread ray-tracing, or using the 
machine as part of a low-latency sound studio, etc), and that what REALLY 
takes the time is updating the config files and otherwise adjusting to 
new versions of whatever packages you just upgraded, which would take the 
SAME amount of time to go thru, for a similar version update, regardless 
of whether the distro is binary-based or from-source.

From a dev and student-dev perspective, more your focus here, it's an 
even MORE important lesson, as these guys will likely be doing a lot more 
package building over their schooling and career than the average user 
(tho maybe not so much more than the average gentoo user), even if they 
don't stick around on gentoo, and the earlier they get the message that 
they can let the computer do what the computer does best and that once 
it's setup properly they don't have to constantly watch it, only check on 
it once in awhile and check the final status and summary messages, the 
more efficient they're going to be as students, devs, and ultimately, 
employees.

> Also, remember that Google's ChromeOS takes a lot of things from Gentoo,
> including the package manager and many ebuilds. The idea here is that it
> has applications in the industry.

Good thought.  Anchor the theory in real-world practicality. =:^)


And since I replied here, I might as well throw in my own suggestion here 
as well.

I don't see any mention of overlays on that list of covered subjects.  
Perhaps mention them under point #2, covering gentoo as a metadistro, 
then cover them in more detail under point #3, delivering the goods.  

Keep in mind that many gentoo devs begin working with project overlays 
well before they're official devs, under the supervision of the devs 
responsible for that project.  Depending on an individual student's 
interest, they could be getting involved with projects and working in the 
respective project overlays before the end of the course, so introducing 
them fairly early on and giving them a chance to work with them, first as 
a user, then as they progress, as a dev, should work out quite well.

Overlays should work in quite well with your coverage of git, too, since 
many of the overlays are git-based.  You could either take Maxim K's 
github idea and do the overlay there, or do something similar using 
overlays.gentoo.org.

Either way you'd presumably work with the hosting provider, github, 
gentoo, or other.  I don't know if you can delete github projects, but if 
I were github, I'd *JUMP* at the chance to cooperate with someone 
offering a college level course integrating github, and would likely 
offer to work with you on integrating anything special you needed, like 
deletable projects, in ordered to do it.  This since once a student's 
using github, they're going to take that practice with them, and in a few 
years, many of them will likely be pushing paid github accounts at 
wherever they end up working.  If you pursue the github idea, I'd 
DEFINITELY recommend contacting github, as they may well be able to 
suggest or even create other features to fit your class as well, and they 
should have a pretty high motivation to do so.  They /might/ even offer 
to sponsor some of the course material cost, etc, especially since they 
can almost certainly write it off as either advertising/PR expense or 
educational donation. =:^)

OTOH if you chose overlays.gentoo.org for your hosting, you should get 
quite close cooperation as well, likely even closer due to scale and 
direct focus, and AFAIK gentoo already has some experience there from the 
Google Summer of Code projects.  But the experience isn't as easily or 
directly useful for students beyond the course and gentoo, as working 
with github would be.  OTOH, working too closely with github might be 
seen as a commercial endorsement you or your school may not wish to make 
(except perhaps if they /do/ actually offer sponsorship, most schools 
know how to make /that/ work), while gentoo as a non-profit that's 
already directly part of the course shouldn't have that issue, and could 
be seen as useful in its own right since it parallels the adaption to an 
existing work environment that an employee (and for that matter, 
volunteer joining a FLOSS project) generally must make, regardless of 
whatever github or etc experience they may have already.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman