Functional requests

[Lord_Farin]

From what I've seen, most problems don't have any PDFs with elegant solutions attached to them. And those that do have such PDFs are more in the easy spectrum. What we need is the opposite: elegant solutions for the hard problems.

So here is my idea: What if, once you solve a problem, you gain access to a thread where the building of such a PDF is made and discussed ? I don't know who made the ones that are already on the site, but if this responsibility is shifted towards the community, it will free the administration from doing it themselves. And we would have a lot more of such HELPFUL PDFs.

One crucial point here: only if/when an administrator marks the PDF as complete/done/good enough, will it be published for the ones who solved the problem to see.

This idea is of course in its infancy, but I believe that it's worth to do it. This way, Project Euler would not only contain problems and answers, it will also contain the solution that gets from the problem to the answer .

Seconded. I imagine some of the hard problems need a PDF of tens of pages to have their solutions adequately and rigorously described; this is more likely the result of a community's effort than of one person.

[euler]

@SoboLAN: I agree entirely with your suggestion as it has always been the intention of the current forums/problem threads on the main website to achieve this purpose: a place for members to post methods, insights, and generic algorithms which ultimately distil the best/optimum approaches. This is why I say, "To ensure that the forum is useful for as many users as possible we need to be aware that not everyone uses/understands the same programming language" at the top of the posting page. Furthermore the idea behind introducing the Kudos system was to help pinpoint the most pertinent/useful posts. However, although the first page generally contains the most insightful posts they all too quickly regress to code snippets in preferred languages.

Having said that I do not want to downplay your suggestion as I am all for it and I know that the team would fully support any ideas which would allow more good quality PDF Overview documents being written. Perhaps we need a push/reminder on the purpose of Kudos and establish a team of members who could take it in turns to assume responsibility for a particular new problem (or older problem) and then to oversee the writing/production of an Overview document? This might not necessarily involve them writing it, but working alongside other members/volunteers in bringing the document towards a completed product.

[SoboLAN]

Well, the Kudos system is a nice idea, but doesn't really work. For example I rarely visit the threads of the problems that I solved, and I very rarely give Kudos. From what I've seen, most members are the same (I base this statement on the fact that on those threads most posts (80%) are with 0 Kudos, exception very old ones). A reminder of what the system is will push it to be more active, but that would last only a few days, 2 weeks tops. After that, people will forget about it again...

I don't have too many solved problems, but I do have most of the ones from the first page solved (at this moment 43 out of 50). And most of those don't have those Overview PDFs (maybe 20 % ?). I suspect that on the following pages the percentage is super low. This is what gave my idea with the members of the community making these PDFs.

Here is how I picture it:

- Maybe (some of) the members should vote which solution is best from the ones already posted in the problem's thread. Or if this would be too complicated, have someone from the staff decide; but let it be someone who knows the problem very well (maybe a member who contributed a lot to that problem ?).
- After the best solution is determined, some members should be chosen for the creation of the PDF. Or maybe ask for volunteers ? The volunteer part would not be very good because of the huge numbers involved. Maybe limit this to members that have at least a certain level ? (Level 4 ?).
- When the people are chosen, the solution should start to be composed by these people. Every one of them can contribute to it. Think Subversion ("rever to previous state" is something very useful here).
- When they think it's good enough, a member from the staff will be contacted in order to verify that the PDF is indeed good. If it's good, it will be published (just like the ones already available). If not, then the PDF should continue to be developed.

Advantages:
- a lot of problems (most of them) will get such PDFs. It's a very important part of the site, in my opinion, one that is currently overlooked
- the staff doesn't have to bother too much with the makings of these PDFs, only verify them from time to time (when the members shout "it's ready" ).

Disadvantages:
- In the beginning, it will require some extra coding and development and organizing the teams. It will be a bit tricky.

[b0rks]

Hi!

I'd like to propose that a member of the Project Euler community could, instead of picking one programming language, pick several of them..
Because even though true, I mainly use python to solve problems, I do sometimes switch to other languages, to either jog memory, or I'm simply continuing from other projects, and not in favor of constantly switching.

I know It's not a big thing, but still a little misleading, to simply state the usage of only one programming language.
I know it might totally either mess up your statistics, or have no effect at all, but a suggestion nevertheless.

All the best,
Borka

[euler]

I know it wasn't exactly the point you made, but here is a reply I gave to a similar request in a different thread:

The point is that it should reflect the language which that member feels is best suited to solving Project Euler problems in general. It is quite possible you start with one language and over time you try out a few more, but when someone gets to, say, having solved 300 problems, and has chosen language_X as his/her preferred language then he/she is doing that based on a lot of experience. If we allowed a per problem choice then it would not reflect necessarily the best language for that problem as much as the language he/she knew at the time they solved the problem. I appreciate that some languages are better suited to some problems than others, and some members might diligently go back and change his/her preferred language for problem_Y, but the idea is to capture a snapshot of the best all-round language in the view of our members.

[b0rks]

Oh.

In that case I do take my suggestion back.. I guess it would make it better to choose the best "all-rounder".

Well, all the best!
Borka

[peedyoJ]

Well.....I would like to suggest/ask if it could be possible to filter problem thread posts by country and/or language.....It would be really great.

[sivakd]

Please use A element instead of JavaScript on DIV element for the list of problems in the "Problems Solved" table in the Progress page.

[euler]

Why?

[sivakd]

If it's a regular link, then it's possible to open it in a tab. Right now that's not possible. Also, it may be beneficial to provide navigation without relying on javascript.

[peedyoJ]

Although no one has responded.....I think the reason for it could be that my previous post might have been misunderstood/misinterpreted.....I meant the thread one gets access to after solving a problem and not the threads on this forum.

[euler]

Sorry, I did see it. I tend not to respond often to "requests" made in this thread but I always take note. Unfortunately the current build does not allow for that easily. Caching became a necessity to reduce server load. The forum pages are literally cached page by page, so, for example, when you use the pagination links to view page X then you are viewing a static, pre-rendered page. They are only re-cached when a new post is made or a post is edited. Generating the forum pages is a surprisingly intensive operation as a number of cross references need to be applied to lookup current member information along with the forum data. Allowing searches/filters would require a series of DB lookups and whereas this type of request is not a problem at most times of the day, when the site is busy a few such requests could significantly reduce website performance. Allowing members to view problems and submit answers is a priority and whereas the forum is valuable it takes second place. I know it's not a perfect solution, but one possibility would be to use a combination of the browser page search and the paginiation links.

Having said that, your request has been made a number of times before so I am still considering other options. One possibility would be giving members an option to display the entire thread without paginiation. This would make the browser search facility far more efficient and reduce significantly the tedium of clicking through page by page.

[euler]

If it's a regular link, then it's possible to open it in a tab. Right now that's not possible. Also, it may be beneficial to provide navigation without relying on javascript.

Actually I can't remember why I used JavaScript in the first place?! I've changed it to hyperlinks, but if anyone can figure out why I might have originally used JS - any unexpected side effects? - then please let me know.

[sivakd]

Thanks euler. Everything seems to be working fine.

[thedoctar]

I have two suggestions:

My first suggestion is whether we can implement a difficulty rating system for problems. Users who solve a problem can then rank the difficulty of the problem, easy medium hard or out of 5 stars etc. and on the problem page it would display the average rating, thus providing useful information on the difficulty of problems. I know that the solved by number can indicate the difficulty of a problem, but later problems, which are solved by fewer people, less accurately indicate difficulty through it. So do people think that the solved by number is good enough, or do people think a ratings system would be useful enough to implement?

Secondly, I was wondering if problems with related concepts could be tagged or colour-coded or linked in some way. Some problems are, e.g. Problem 114 (View Problem) and Problem 115 (View Problem), yet others like Problem 137 (View Problem) and Problem 140 (View Problem) are not. And maybe all problems about primes, or Pythagorean triples could have a different font colour on the progress page?

[rayfil]

Any difficulty rating system would be entirely meaningless, even if problems were ranked by those who solved it. Some people even have difficulty with Problem #1. How would they rank it??? Alternatively, some people have solved over 300 problems within a few weeks. How would they rank them all???

Any of the problems can be considered "easy" if your math knowledge is good enough to know exactly how to approach it. It then becomes a simple matter of writing some code. Rating the difficulty of a problem is VERY VERY VERY subjective. Even the newer problems which may have been considered as "easy" by the dev. team would have fallen in the very difficult category if compared to the first 150-200 problems.

[TripleM]

Also, figuring out that the problem is about, for example, pythagorean triples is often one of the biggest first steps in solving the problem, so in the majority of cases you could not use such labels without giving a lot away.

[thedoctar]

Any difficulty rating system would be entirely meaningless, even if problems were ranked by those who solved it. Some people even have difficulty with Problem #1. How would they rank it??? Alternatively, some people have solved over 300 problems within a few weeks. How would they rank them all???

Any of the problems can be considered "easy" if your math knowledge is good enough to know exactly how to approach it. It then becomes a simple matter of writing some code. Rating the difficulty of a problem is VERY VERY VERY subjective. Even the newer problems which may have been considered as "easy" by the dev. team would have fallen in the very difficult category if compared to the first 150-200 problems.

Yes. Obviously people would find each problems to be of different difficulty, but as enough people vote on each problem, the rating would reflect the 'average' Project Euler user. And the problems are "easy" if your maths knowledge is good enough, but this would help those who found them hard, since, if the average rating was easy but you found the problem difficult, then you would know that there was some part of mathematics that you didn't know or understand.

Also, figuring out that the problem is about, for example, pythagorean triples is often one of the biggest first steps in solving the problem, so in the majority of cases you could not use such labels without giving a lot away.

It's true that some problems have hidden concepts part of the problem solving process, and you wouldn't want to reveal that, BUT there are some problems, like 137 and 140, which are obviously closely related. And when a problem asks to find Pythagorean triples, or primes with certain properties, then it's obviously about Pythagorean triples or primes, as the problem states so itself. It's much more fun, at least I find, solving related problems and I learn a lot more about the concepts in related problems, as the problems get harder, the easier earlier problems give me an introduction into the ideas of the concepts and the harder problems force me to research more complex ideas about the concepts in the problems.

[hk]

Yes. Obviously people would find each problems to be of different difficulty, but as enough people vote on each problem, the rating would reflect the 'average' Project Euler user. And the problems are "easy" if your maths knowledge is good enough, but this would help those who found them hard, since, if the average rating was easy but you found the problem difficult, then you would know that there was some part of mathematics that you didn't know or understand.

There's one caveat.
Obviously one can only estimate a problem's difficulty correctly after having solved it.
That would mean that the population that voted for a problem's difficulty isn't representative for 'the average Project Euler user'.

[thedoctar]

There's one caveat.
Obviously one can only estimate a problem's difficulty correctly after having solved it.
That would mean that the population that voted for a problem's difficulty isn't representative for 'the average Project Euler user'.

Hmm, I hadn't thought of that! Oh, well... I wish there were some way of having some accurate representation of a problem's relative difficulty.