Locked/Archived Threads @ Project Euler

[0x666]

I know thread locks is needed for decreasing CPU usage for project euler site, but...
Think of it, that MOST submits/solutions will be done for particulary OLD problems (because most easiest problems is OLD).
So acctually these locks means that MOST users can`t post. It is not fair. You would rather think of other ways of decreasing CPU usage...

[euler]

Although CPU usage was a serious issue with our previous hosts, and certainly was the original cause of the threads being locked, our new hosts (pcsmarthosting.com) have solved that problem by providing managed dedicated hosting for us. So the issue is not so much CPU usage as much as there is now little chance that anything else can be said in those longer threads.

Once a thread gets to a certain length there is a decreasing chance that a member will read all the preceding posts, so they will begin their post with something like, "I'm not sure if this has been said already, but...". That member has now increased the number of posts by one and consequently decreased the likelihood of the next member reading the preceding posts. This process continues and the longer the thread the more likely it has been said before and the less likely that someone will notice.

I know that the thread for problem one is an extreme case, but at the time of making this post there are 949 posts spanning 38 pages. Do you honestly think that there is anything new someone can say on this particular problem?

In one sense you are right. It is not fair that newer members will not have the opportunity to post in the threads for old problems and there are some problems with open threads that some members new to these types of problems will currently find too difficult to solve. But the whole point of Project Euler is, by starting with the easier problems, to develop skills and learn new approaches that will ultimately help access the previously more challenging problems. The threads are a perfect place to learn. And if these members are unable to solve the more difficult problems then this is the stage they are at: they are still learning. Would you allow students at school to tell the teachers how to teach? But as these "students" graduate they will become teachers themselves and it is likely they will now have gained sufficient skills to access the more challenging problems and then will have opportunity to post their insights; if that is their goal.

Of course, these points do not address new members who are already skilled problem solvers. But if they are at that stage they will take a quick look at the first few pages of the locked thread and realise that anything they were going to say has already been said in one form or another. Also one of the issues the team are trying to address is a balance in the difficulty of new problems. The aim is to rotate new problems on a sliding scale from easy to hard. So once again there will be opportunity for newer members to post.

The bottom line is that we want the threads to be a useful resource and we didn't want threads to become so long that members are not able to make full use of all the posts made. To have such long threads that were unlikely to ever be read is no use to anyone and is almost insulting to the people that have taken time early on to outline their methodology.

I understand how satisfying and rewarding it has been to have solved a problem and then have an opportunity to celebrate that by sharing your method. But I hope you understand the reasons behind this decision. If anyone can suggest a better solution we will gladly consider it. This thread is not locked!

[joshbowman205]

Perhaps a fixed cut off limit, to avoid any bias. For example once 100 posts are reached or 1 month elapses the thread is locked automatically?

The thread system is not ideal.. im not sure what the current consensus is.

[rlindley]

I don't see that the locked threads are a problem. They are still readable. If a user studied the thread and thought he had something truely new and interesting, then he can always post it over here in this forum. Then if it is just BS an admin can PM him with a nice note and then delete the thread on this forum. I think any serious solver would figure that out if he thought he had made a new discovery.

[edit] Again, I did not say exactly what I was thinking. Of course a user should not post a solution over here, but he could post (or PM) that he has something new. But if he did post a solution, admins must be on guard for that sort of thing all the time anyway.

[euler]

You're absolutely right. If a member PMs one of the team with a remarkable insight into the problem that has not been said before then the thread could be temporarily unlocked to allow a new post to be made.

For your information, the difference between a locked thread and an archived locked thread is that locked threads cannot be modified (edited or posted to) by members, but they can be modified by the team. The posts are cached in the same way as unlocked threads. Archived threads, on the other hand, cannot even be edited by the team as fully parsed versions of the posts are cached to speed up displaying them, which means that all the custom tags have been converted into HTML. However, the team can restore them to unlocked status (as the database still contains the orignial posts), which will delete the archived cached files and produce regular cached files.

[hk]

Perhaps a fixed cut off limit, to avoid any bias. For example once 100 posts are reached or 1 month elapses the thread is locked automatically?[/b]

Currently we locked threads with 100 or more posts. As Colin explained from the user point of view there is no difference between locked and locked(Archived).
Colin made a nice overview in our admin panel where we can see how many posts there are and where we can lock/unlock/archive the forum for individual problems.
I think we can handle this manually.

The thread system is not ideal.. im not sure what the current consensus is.

Sorry, I do not understand, what do you mean?

[joshbowman205]

The thread system is not ideal.. im not sure what the current consensus is.

Sorry, I do not understand, what do you mean?

I mean that for someone to gain an understanding of the problem from the forum they have to read a lot of repetition, slight variations and timings even in the first few pages. Sometimes good explanations follow later on..

[hk]

Reading through through a great many posts to get the idea is certainly not ideal.
At the moment we are discussing in the team to set up a system that enables us to publish an overview for a certain number of problems that is accesible when one has solved the problem.
A crude version of that system is already running.
To demonstrate it I made an overview for problem 1.
If you go to the problems page and look after problem 1 you will see this icon

icon_resources.png

If you click on that you are taken to a directory for this problem where one can view a pdf with an overview.
You will also find a file that's the editable source of this overview.
It's placed there so that we in case of necessity can retrieve it, edit it, make a new pdf and upload both again.

Mind that this is a very raw idea about the system that we are discussing. Perhaps we do change the interface. But something like this will certainly answer your concern.

[joshbowman205]

That looks good. Just need to replace the .rtf with .tex?

I wonder if some kind of suggestions box might be appropriate? ( for changes to the pdf for that problem)

[hk]

That looks good. Just need to replace the .rtf with .tex?

As anyone can view the pdf that seems not necessary. That one is for general purposes.
The editable source is there for team purposes.
We are still discussing the final structure.
Up till now we think we can manage .doc, .tex and .rtf. as authoring tools. Just what the author prefers.
Same stance as we take regarding the use of programming languages.

I wonder if some kind of suggestions box might be appropriate? ( for changes to the pdf for that problem)

As said the system is still under discussion.
But I think a system like this will reduce the need wading through endlessly repeating forum posts.

[kinyel]

hello,

The issue is not so much CPU usage as much as there is now little chance that anything else can be said in those longer threads...
...
...But I hope you understand the reasons behind this decision. If anyone can suggest a better solution we will gladly consider it. This thread is not locked!

I not have really a solution, but perhaps an idea :
classify in differents threads for a problem :
* by type of solution (for example brute force, formula...)
* by language or category of language (solve with mathematica or maple, and solve with C or php, and solve with assembler, etc. are very different ... and solutions with "esoteric" - for me - languages are too strange for learn math or algorithms with them

the result of classify : to choose a language clear for me
or to choose a type of solution I've not think ...

[3n1gm4]

Hi, I'm new here, I pass some of my free time here for fun, sometime I've found my exact code already posted, but sometime my solution is a bit different, i.e. I often use python generators and I don't see their use in most of the python solutions I saw.

I think classification would be useful. I'd make the user choose his language for each solution too, someone uses different languages depending on the problem.

What's the problem keeping threads opened? Space in the db? CPU used to query the db? You use a cache system, don't you?

[euler]

I think your post is a case in point as to why we lock longer threads on the main website. This thread was only eleven posts long when you asked your question and you didn't notice the second post (that I made), which answers it.

[3n1gm4]

I think your post is a case in point as to why we lock longer threads on the main website. This thread was only eleven posts long when you asked your question and you didn't notice the second post (that I made), which answers it.

Should I send to you what I think is a good and alternative solution?

[euler]

Certainly if you genuinely feel that you've got something different then PM me. However, please be aware that there are two ways in which it would be considered uniquely different and worth adding: (i) it uses a completely different algorithm not already mentioned, or, (ii) it makes clever use of a library in the member's preferred language to reduce the amount of code necessary.

Given the lengths of locked threads, reason (i) is highly unlikely, but never impossible. Reason (ii) is the most likely, but the problem then is that none of the team may know the programming language well enough to make that call. Yet if you feel that what you've got would be useful then PM me, outlining very briefly what your post would offer.

[aantthony]

Maybe a poll like box, where you can choose your way of doing it, like maybe the language, or time it took, or method you used. Then you could see what most people did.

[stijn263]

That'd be cool. But I'm afraid you'll get the following response from one of the teammembers soon:

Currently we are not able to handle any functional requests that imply changing the functioning of the website or this forum.
Sorry.

[hk]

Aye

[Groxx]

Though I miss being able to post, I must admit that 99% of what can be said has been said.

Maybe a locking mechanism that requires a user to be X days old, and have achieved level 1 or 2? That'd prevent newbies and cheaters from gaining immediate access to the ability to spam, while allowing the threads to still be updated. Theoretically, they'd understand the community well enough after that point to contribute, instead of just adding to the noise.
In another case, say someone uses a new/fun language, like LOLCODE (if possible, not sure on its status any more) or Brainf**k (yes, they're both 'joke' languages, but you get my point), and wants to post their solution? Or finds a new algorithm to do X even faster? Or screw efficiency, maybe someone used a genetic algorithm to get the answer. That could be interesting to see.
Even though it may arguably "improve" the discussion, very few will even consider PMing to get a post made. How many have been requested thus far?

That said, I really really like the PDF links for a detailed explanation. They're far from complete, and I unfortunately can't contribute very much to creating them, but that quite possibly stands to "solve" the problem.

[daniel.is.fischer]

Or finds a new algorithm to do X even faster? Or screw efficiency, maybe someone used a genetic algorithm to get the answer. That could be interesting to see.
Even though it may arguably "improve" the discussion, very few will even consider PMing to get a post made. How many have been requested thus far?

I remember one request.
It may be that people don't bother to ask, which would be a pity, we'd be happy to temporarily unlock threads for interesting contributions. So if anybody thinks (s)he has one to make, please check if somebody else already made it, and if not, tell us about it.
But it may also be the case that those who find the most efficient or otherwise interesting ways are usually the first to solve the problems (of course, should Donald Knuth decide to register tomorrow, ...), so the good points have already been made.

That said, I really really like the PDF links for a detailed explanation.

They're far from complete,

Yes, it's surprisingly much work to write one, and we have so little time

and I unfortunately can't contribute very much to creating them, but that quite possibly stands to "solve" the problem.

But you and others could contribute by giving feedback. What needs to be better explained, what is good? Are they too short/concise, too long/verbose?
For which problems would overviews be most desired?