There will always be projects that are left out. I can't understand why such a great initiative is used for another round of Google-bashing. Google is spending loads of cash to help open source projects, and still some people have to protest.

There are plenty of FAQs that explain the rules, and of course you need to present your project well for Google to accept it.

In any case, the rockbox project (www.rockbox.org) was happy with the results from last year and is looking forward to the goodies we'll get this year.

Peter (gsoc mentor for rockbox)

So two years ago we registered 10 projects for GSOC, received plenty of applications from students and diligently reviewed, interviewed and finally picked the best applicants.

Half the applicants failed to reach the halfway stage, with some producing no more than a few lines of code. Only two collected their final payments and only because we were generous. Of ten projects only one was salvaged and needed extensive re-writes before it was usable.

To us, it proved that students lack the discipline to work outside a school or office environment. Working at home with all the distractions that brings, while the rest of their friends were enjoying summer holidays seemed too hard for most.

The mentors tried coaxing the students into working harder but it didn't work. Mentors were achieving lower output themselves while they tried to help their students and this defeated the whole purpose of GSOC for us. Instead of gaining ten new developers to our team, we lost existing ones to babysitting duties.

Will we ever participate in GSOC again? Probably not. At least not while the only valid applicants are students. If Google alters the scheme to accept experienced developers and existing project contributors then we would definitely reconsider.

You make it sound as though XFCE is exclusive to Xubuntu. It isn't. In fact, it was around long before Ubuntu. It's a shame it wasn't accepted anyway. It's a great little desktop but could use some love in places.

why linux need 25 different desktops ? gnome kde xfce , this and that ..

Finish the first one ! Once that one is bug free you can then focus on writing some user applications like a photoshop or paintshop ( please spare me the Gimp ..that's a toy compared to Photoshop) and Premiere clone. Or a solidworks clone or an Orcad clone.

And then fuss around with the eye-candy.

I can explain: it's because there is no "the linux", the desktops are not bound to "the linux", finishing a desktop is like achiving victory in Iraq, people like to generate a new, fresher wheel from time to time and anyway it's often easier to start anew than to convince Software Badr Brigades to accept and integrate your ideas. Why was Gnome created again, hmmm?

Consider that "\"the\" Windows desktop" for example is also asking for a crispy new 200 dollar bill every few years.

"Or a solidworks clone or an Orcad clone."

One of your wishes has been granted some time ago. The solidworks for linux clone is called Pro/Engineer. The closest that you can get to Orcad is gEDA, I don't know how close it is because I leave that to the sparkys.

A bug-free desktop?

What fantasy world does that exist in?

> To us, it proved that students lack the discipline to work outside a school or office environment.

Nonsense. Early versions of linux [amongst a myriad of other examples] will quickly disprove this generalisation. There are plenty of students [whether compsci related or not] chucking good quality code at open source projects or developing their own project.

Some don't do well working in this particular way, without a doubt, but it has sweet fanny adams to do with being a student or not.

The more obvious "proof" is, you need to pick someone who will deliver and you didn't. See what /you/ got wrong yet?

Your project failed to pick good applicants, from bad ones. That's not easy if you don't have an existing bunch of a code from them. But, if you had the code you wouldn't need to pick the applicants :)

This is why open source "show me the code" works so well, because it's easy to pick who would deliver good code after they have delivered it. It's a 20-20 hindsight system :)

But perhaps you were "diligent" when it came to picking someone who perhaps looked potentially skilled at coding and knowledgeable about the project, without considering the more obvious necessity. The one that failed in your case : Would they actually deliver anything.

This is your failure though, and you should learn from it. Giving in on GSOC and blaming it on some sweeping generalisation about students and summer holidays just shouts even louder that you lack the skills, were at fault and haven't even the experience to notice that's the case.

Perhaps this is the downside with open source style "show me the code" stuff, the project contributors / leaders neither need to have, nor need to develop any kind of skill picking or managing applicants a priori.

So yeah, that doesn't bode well for GSOC, perhaps google would be better - if their motivation is to add to existing projects rather than simply pay existing developers, with a GSOManagement ?

There's always concern and debate (ok, visceral hatred and dung-flinging) inside the Debian project. If you were Debianizing a bikeshed there'd be a month-long policy discussion about what colour to paint it followed by a violent disagreement about the creative licensing position of a paint job, whether tricycles should be supported, and the localisation of the word "colour". Not much time left for actually coding your VGSOC submission when there's a policy flamewar to join!

what's it like living in the soviet union? or do you work for microsoft?

"The solidworks for linux clone is called Pro/Engineer."

Which is also availiable for windows. Whilst a much better program (IMO) its ease of use is way down the scale.

"why linux need 25 different desktops"

Because people like choice.

"Nonsense. Early versions of linux [amongst a myriad of other examples] will quickly disprove this generalisation. There are plenty of students [whether compsci related or not] chucking good quality code at open source projects or developing their own project."

Your logic is faulty. Just because all crows are black does not mean that everything black is a crow.

All the previous GSoC attempts that I have been involved with/know of have been somewhere between 'not very useful' and 'actively counterproductive'.

The OP was not (as far as i can decipher) suggesting that students are nececarily bad at contributing, he was saying htat GSoC students are. If they won't do it without money, theres a pretty good chance they won't do it with money either.

Debian is completely open, meaning all debates and disagreements are public, has very firm moral guidelines, and a very clear understanding of the meaning of "free". This naturally causes tensions with people corrupted by society's norms.

If only countries/the world could be run like Debian.

Tux because where's the whirling spirit?

There is no reason why all sponsorships for "summer of code" type work need to come from Google. For example, the GNOME Foundation has sponsored projects in past years to supplement those sponsored by Google.

If Sun thinks the program is valuable, then they could sponsor a few students to work on OpenSolaris projects. HP has been a big supporter of Debian: they might want to do the same for that project.

Google should be applauded for sponsoring the students they have, rather than disparaged for the the ones they declined.

Living in Soviet Union was, actually, rather good, contrary to the propaganda

"If you were Debianizing a bikeshed"

The concept of bikeshedding (as described by C. Northcote Parkinson) is in no way unique to Debian. It's been a feature of every software project I've participated in - both open and closed source - where there have been multiple developers.

For example this particular GSOC project, for FreeBSD:

http://www.freebsd.org/projects/ideas/index.html#p-overhaul-config

If they won't do something similar, at some point, without money, they are unlikely to do it - true.

On the other hand, that's a whole world away from doing that precise project, at a point when the weather is good, and they could be less nerdy and go out and enjoy themselves. Open source generally shows various boring jobs tend not to be done, unless people are motivated by factors like hard cash.

"Living in Soviet Union was, actually, rather good, contrary to the propaganda" ..... By Maksim Posted Friday 28th March 2008 10:33 GMT

Those Ukraine girls really knock me out, Maksim. God Bless Mother Russia. Social Communism at the XSSXXXXual Level is a Dish best Servered Hot.

It is certainly Capable of Stripping Exclusive Capitalism Bare to the Skull and Bones.

And the Jolly Roger because it is entirely Appropriate?!. :-)

I'm sure those dorks at Google like nothing better than spending the nicest part of the year locked in their basements waiting for compiler runs to finish. And I'm equally sure they're baffled why nobody else wants to. How about Google's Winter of Code? Or are the students too busy getting wrecked over Christmas/Hannukah to churn out some lines?

"To us, it proved that students lack the discipline to work outside a school or office environment. Working at home with all the distractions that brings, while the rest of their friends were enjoying summer holidays seemed too hard for most."

Woah, woah, hold the phone. As a student about to graduate, I did my year in industry last year. Now and then I worked from home, when I got just as much done. I work far better on my own rig, at my own desk, with my own music. Fortunately, the place I was working let me have a similar setup in the office, which kicked arse. In any event, I really resent hearing this "lazy student" bullshit I keep hearing. I haven't slept properly in years because I have Uni work to do. When I'm not doing that, I'm doing freelance work and going to a part time job, so that I can afford said University education (cheers Tony).

I've gone off on a slight tangent there. Apologies. Anyway, my point is that not all students are lazy buggers. If they are, sack 'em, or improve your interview techniques to weed out the useless.

Hey! I never used the word lazy!

I was a student once too. When I said that "students lack discipline" I obviously wasn't talking about them all but those who we accepted to projects. I can see how you might have taken what I wrote to mean that all students are lazy, but that wasn't the case.

I take strong issue with the suggestion that we didn't know good the good applicants from the bad. The selections were made by four career programmers, all experienced in hiring staff. We selected the best ten applicants from the ~110 who applied, there is no question about that. They were picked not only for their experience and qualifications but primarily for the effort they put in during the application process. Half failed to even submit a detailed application.

Many good students still spend their spare time writing code for Open Source projects, including ours, but at least in 2006 these people did not apply through GSoC. Had they done so they would have been selected.

Are all applicants to GSoC useless? Of course not, however it does seem that the biggest names attract the cream of the crop. Just what percentage of GSoC projects are completed? I'd be curious to know if our experience was the exception or the rule.

If Google were serious about investing in open source they'd open up SoC to anyone and not just students.

Well, you Gates slaves stick with your bug-riddled crash-tastic Windows set-up if you like, we'll have the last laugh.

Paris, because she sucks like you

You sound like a very diligent student, and I salute you, but come on, any of us who have been to uni know that you are the exception rather than the rule.

Hell, I personally spent a solid majority of my time at uni drinking, shagging, and engaging in experimentation with various recreational substances. Fortunately for the world at large, I spent my downtime writing code, and learned to do so to a pretty high standard.

That's not as inspiring, but is far more typical of the university experience in the majority of cases.

There's a grain of truth to most stereotypes.

Always best to get it out of your System in Order to be able to move on to Greater more Focussing/Focussed .........Things/Thoughts to Energise.

Although there's no reason for you to deny yourself anything pleasurable in its Moderation.:-)

GSOC is only worth bothering with if it leads to a job, or it's not about money. Google is using it, in part, to identify people to recruit.

Therefore, people will aim at the more popular projects, and won't bother with the piddly ones.

The students have three months to write the code in, it is to be presumed that the task isn't trivial. They are paid 4500USD.

I don't know what temporary work pays in the US, but in the UK 4500USD over three months is not far from the national minimum legal wage (i.e. for crappy jobs). Temporary work from agencies tends to pay higher than the minimum wage. Coding work should be priced considerably higher.

So, if it's not going to lead to a job, you're asking them to give up the best months of the year, doing something nerdy, missing out on exercise, holidays, alcohol, sex etc for poor pay. Wow, that sounds *SO* attractive! Your only hope is that they're contributing to something world changing, or something that is a Calling rather than a job.

(Actually, it's even worse than I mentioned, because in total it spreads over more than three months)

if its over 3 months , and they work a 40 week that's $9.37 hr 4 months $7.01

You can flip burgers out here for $10 / hr /.

Oh minimum wage in the city of SF is $9.36 /hr

The problem with getting people to finish a project is one I've dealt with myself (uni project where my "team" mostly solely spoke mandarin, apart from one who spoke cantonese and one who spoke reasonable english.....what happened....one went back to HK and didnt come back......2 sat back and smoked cigarrettes the whole summer and pled they couldnt understand the project....leaving me with minor help from the one team member who spoke slight english to write a 6 person programming project.....yeah that didnt get finished....didnt even get credit due to the lack of effort from others.....)

On another note

As a 2004 computing grad who is still stuck in retail (as well as most others who graduated the same year or who did another higher level (most of whom were and are damn good coders/hci etc)) I'm always amused by the "severe lack of computing graduates" "IT suffers dearth of staff" headlines.....

If companies are so short staffed then why are none willing to employ anyone with less than a first or rarely an upper second???

or even consider someone without 3 years commercial experience?

That really irks me, especially when I see some of the dross being put out as "finalised code" (its something called a deadline you know >> reply ive heard several times when I've commented about the poor reliability and bugs in an app) or "completed machines" (I'm primarily a hardware geek / HCI nerd to the extent people cant fathom why I'm stuck in retail and the "service engineer" they get has the engineers job.....usually as they memorised the answer key used by their "new university")

And no one mention university "careers" services....that lot are hopeless....brochure distributors more like.......

and people wonder why I'm considering retraining in a trade...might be some hope of finding work in the field I studied for. plus the course might cost less than £5000

> I take strong issue with the suggestion that we didn't know good the good

> applicants from the bad. The selections were made by four career

> programmers, all experienced in hiring staff.

Well, just listen to yourself. "Career programmers" - WTF does programming, career or not, have to do with picking the right people for the job?

Yes, perhaps as a programmer you picked people who had the programming skill to complete the project - I said as much...and yes, a halfwit could reject applications that weren't detailed enough, so well done there.

But, why didn't you reject the 10 candidates you picked? Don't tell me, sit down with the other 3 and beat yourselves up about it. Don't blame the candidate, had it been a corporate job they'd have been laughing on your bosses dollar...and you'd have had someone else asking you the same question.

But, the facts speak for themselves, the people you picked failed to deliver.

Only a complete buffoon would argue at that point that his team picked the right people and start willy-waving his teams' recruitment skills.

Don't get me wrong, as I said, its not a skill that open source requires that much, and, if your "career" programming involves something outside OS, working for someone else, they'll probably have folk higher up who will worry about these things for you....and everyone hires a few turkeys, and everyone blames someone else for something they did from time to time as well....

Rejection means potential new features for OpenSolaris will not benefit from the GSOH make over.

Was the Good Sense of Humour acronym used on purpose, or were you perhaps writing 2 articles at the same time ;~)

I was going to apply as a student to GSoC, primarily because I am dissatisfied with my current contract programming arrangement (primarily their ability to pay me on time).

In the hours leading up to my clicking the "submit" button on my application, I was offered not one, not two, but THREE job offers making a damn sight more than $4500.

The good students - the ones with experience and self-motivation, are already working. In fact, since GSoC is supposed to be treated as a full time job, that cuts in even more to their potential base, because it excludes everyone already working ANY job, from a grocery store to a freight company from participating, because if they quit their job to participate in GSoC, they aren't likely to get it back in the fall. Only the unemployed kids, who you'll find are typically the least motivated and most likely to spend the summer on drugs, sex and rock and roll, are therefore eligible.

In order for GSoC to be attractive to the right set of students, the stipend has to AT LEAST double (or the expected workload halve) and some consideration be put into the fact that students will often have bills to pay, and thus waiting until July to get paid any substantial sum is out of the question.

Michael, I think you're the one missing the point. How was it that 10 bad applicants were picked? With the obvious worst candidates weeded out, then statistically to pick 10 bad apples would be pretty hard unless the whole barrel was rotten. One interviewer might be spectacularly bad, but all four?

I should point out at this point that I wasn't one of the four mentors who made the selections, pointing out their qualifications and defending their choices had nothing to do with my own ego.

Would these applicants have been selected to work for us in our own companies, probably not. Projects involved in GSoC can't be that picky, they can't decide to hold out for better applicants, promote from within or go head hunting. You have a fixed number of applicants, you select the best of the bunch and you cross your fingers. Of course we could have decided not to go ahead but we took the chance that these students would help progress the project and we couldn't have predict the absolute failure that resulted.

As someone has already pointed out, the good students either already work in better paid summer jobs or they are working on open source projects already (unpaid). Those that are left will gravitate to the big name projects leaving those projects which really need the input of extra developers with the dregs. The closed pool of applicants is the biggest problem with GSoC and the reason it failed for us and others. Change the rules and allow non-student or existing OS developers to be involved and GSoC will be more than a cheap PR excersise.

>> Rejection means potential new features for OpenSolaris will not benefit from the GSOH make over.

GSOH? Are you trying to say that OpenSolaris is fat?

Bashing GSoC is like complaining about someone who is spending their own money to buy shoes for the homeless then complaining that they are not also buying boots, socks ...

What GSoC is doing is a great things.

What would also be great is a highly organized mentoring program for non-student programmers. Some place where programmer who are *serious* about getting involved in open source can be matched with projects *serious* about helping them get started.

Someone should be doing this, but Google isn't obligated.