==/!\ This message may contain some errors or approximations ; it's a quick answer to give my sentiments.==
==
I don't think publish solution is a good idea.
I think we can't stop people who want publish, but we can forbid links in signatures, moto, to solutions blog...
==
But democracy will speak, and if it would happen that publishing is allowed or promoted, then, if it's spoj_inside, here my 'recommendations' :
* Only after solved a problem we should be able to see editorials, not before, else solve a task is a non sens.
* The psetter "solution" should strongly be the first to be "visible", it could be a part of rule for setting a problem : to publish an official solution by psetter. (Visible by all when the problem is first solved)
* Other "solutions" could be published in a kind of forum thread like there is at Project_Euler. (PE is a great reference!!!)
* It is important to give rules about license on those publishing ; I strongly vote for Creative Commons ones that could be enhanced by other in the spirit of author, completed, ... and such that no commercial use can be made with.
* Anyway I prefer not to see that!
==
Publishing solutions will kill classical as it is one of the only remaining coding place where you can find lot of (easy or moderate or hard) problems without any available solution so that you can search on those that suit the best for you and solve it and can be proud of solving it by your own.
In contrast Project_Euler (excellent quality) have very few problems with non available solutions, and those ones are often way too hard for me (or I suppose that), I really would like no publishers have spoiled them. I remember the time where I struggled to solve my first ones and I saw some known people massively cheat... The good answer of Project_Euler was to give credits only to solvers who solved fast recent tasks : excellent answer ; but way too hard for me. The funny part was the Eulerians absent from new ranking table!
After I found in SPOJ a way to practice my poor skills, and I tried (very often) to solve tasks with a very fast method so that mine was truly mine, and to continuously experiment new ideas : it was a vector of progress for me. I never progress as much as when I tried to time like numerix or Mitch ; many thanks to them by the way. (I hope some users will follow some similar experiments to understand by their-self so many coding tricks). There's maybe other reasons whom I'm not sure, and I'm not the only psolver to do that, each of us have its own reasons maybe different.
For me a problem with an available solution that can be googled have (and it's a shame) no value at my eyes.
I'm proud when I can solve a new uspoiled task, and I feel bad when a beautiful problem fall by the fault of an irresponsible publisher.
It is maybe rude to qualify them of irresponsible, but I just note that I don't think any high ranked user have published his solutions ; we massively find middle skilled psolver. kokosek is maybe the only known exception of a talented psolver that publish some solutions.
I think the only way to learn something is to search methods by our own, read unrelated papers for general methods.
Read a solution won't profit to anybody. If you think you need that : just go to codechef tutorials or other sites (I don't know them, but there's maybe a lot of similar as pointed previously).
If SPOJ should become a place with published solutions, I will maybe solve no more tasks, and only work on setting new problems. (I yet work more on setting new tasks than solve others, but I like both, and I've too a full time job unrelated to coding.)
Edit : There's challenge section too with lot of good remaining tasks, Mitch is the boss here and I understand well why it is a good investment.
(and I write 'no value', it's too hard word ; sorry).