Algorithms Forum?

I remember a discussion about whether there should be a forum for algorithms at DevShed. Afair ppl agreed that it would make sense.

What has become of it?

I assume that all (or most) ppl who supported this idea haven't changed their mind, so I'd like to clear the language matter with this poll.

*edit*:
If you voted for "Other", plz tell us which one!

I favor the opening of the algorithm forum and was actually gonna open a new thread to ask the same your are asking. I ain't that big of a fan of it, but also intrested in exploring the topic. So how about it, is it coming or are we just hoping it´ll come?

yeah..it's been brought up a couple times to Devshed. Just wish they actually paid attention to what goes on in the forums...

---John Holmes...

Hi

I've been asking for the algorithms forum, but as I asked it 3 times without getting any response from the moderators (saying "ok, we'll open it" or "it's silly, forget about it"), I finally got tired of asking for it.

And now I see a C forum. How much time the member which had the idea lasted to get it ? How many posts he needed ?

Well, I know that any kind of cross-posting is not very popular, but I think that if we asked for it in the newly created forum, the one that administrators are paying attention of now, maybe we'll get a clear response. If you agree with me, please go to the C forum and sign.

Thanx !!

BTW: I voted pseudocode.

What good will an algorithms forum do? I have seen maybe one or two threads about algorithms in the PHP forum the past 6 months, and that's *the* highest volume forum on this site.

I've seen more threads in PHP alone (and I don't have time to read all threads). Many questions are of a "how to do this" way and can be answered by posting simple algorithms instead of pages of code. And perhaps such questions just don't get asked because there is no forum for them. Anyway, it's not the traffic, it's the quality that determines the good of a forum.

In most cases, you get *lots* of algorithms - some standard - some ad-hoc - everywhere in books, websites,etc.

Usually (at many other forums that I visit), people ask questions about how to implement a particular algorithm in their C project, or PHP project, or Perl program. This effectively constitutes a post in the respective forums for C, PHP or Perl.

If you are planning to discuss Cormen's book algorithms in detail, or maybe inventing a new algorithm by deep discussions, I'm up for the Algorithms forum.

Right, so why not here? It's more comfortable to do something if you don't have to switch communities to get help.

Yes, but ppl are also asking questions that are answered better (more generally applicable) by an algorithm. E. g., the traditional "how do I make a [PHP] program that prints a table with photos" question can be answered by code or an algorithm (or STFW, of course). The first way saves time for the asking person, but the second gives that person (after having been coded) a much better overview over and understanding of the program. Since many ppl (including me) are here to learn, the latter way is quite attractive for them. Others can still get their implementations (or STFWs) in the respective forums.


I haven't read the book. But creating new algorithms could be also fun.

I consider algorithms to be solutions to more complex and time critical things than trivial stuff like "how do I make a [PHP] program that prints a table with photos". (Actually, this question is so vague that to give a solution is almost impossible, let alone an algorithm!)

When I think of algorithms, I think of generic solutions to recurring problems in the field of computer science, problems that can normally be solved in exactly the same way regardless of programming language. There are few problems in PHP that match this.

PHP is a highly specialized tool geared towards the web, when people ask for "how to do x", they normally ask how to write the application that solves x, not for an algorithm.

Again, the number of discussions about algoritms is minimal, there was one yesterday however, regarding how to find a prime number. In this case, a user posted an algorithm written in C, and this algorithm is clearly directly translatable to PHP. This rarely holds for the problems posted in the PHP forum, I shan't claim to know what is the case for the other forums.

I don't pretend to think that I would find much use for the forum. I would rather be observing and reading what other discuss, but in general the algorithms facinate me. I agree with andnaess, that algorithms are solutions for complex tasks that can be done all over again, as long as the objective is the same, regardless of the system we are working on.

About PHP and algorithms, it has come clear to me that they do not match. Like I was told, in a discussions about the same subject in the PHP forum, PHP already contain many known and fast algorithms to accomblish tasks and it would be more like waste of time to try implement algorithms to PHP scripts. But I think C/C++ and other programming languages are a different story. Often I see many books, in various bookstores, that discuss algorithms specificly. And even books, about programming languages (Delphy, C/C++, Java) contain chapters about algorithms. One might think that algorithms are not a rare matter to be discussed. But this is just my opinion.

Datamike, andnaess:
I agree with you in most points. My example was indeed bad (I just couldn't think of anything else at that moment).
About PHP: it does have many things already built-in, but doesn't have others. Those could be worth discussing.
Algorithms: can be imho solutions for any kind of problems, completely independant of computers (languages) and/or of how generic or complex they are. They can even be trivial, like some sort() stuff.

Indeed, algorithms can be many different things and be very simple. Sorting algorithms are fairly complex, the analysis of shellsort for example is a very complex task (shellsort in itself is not hard to code though).

Of course, the study of algorithms is essential to anyone studying CS, and most schools have a course dedicated to algorithms (typically in combination with datastructures) very early on.

Algorithms are very interesting, but in order to have a meaningful discussion about them you need to have a very firm grasp on both programming and statistics (and probability). The study of algorithms today is mostly done by people with a very good background in mathematics, and the problems found in this specific area of CS are extremely esoteric. (Crack the issue of NP completeness and there's a million dollars waiting for you).

So what I'm wondering is what would be discussed in such a forum?

Certainly not the $1.000.000 problems

But there are other questions worth discussing.

Also, the forum doesn't need to contain only high-level CS. It could be a place for those who "just can't find a way of doing that". You probably have yourself experienced blackouts when you couldn't solve a trivial-looking problem.

And there might be questions about how to make a solution more efficient, in the sense of algorithms (not compiler options etc.)

Probably a lot of regular expression matching and string matching in files could be discussed as algorithms.

How do I match all of the urls in my text file?

How do I verify an email address?

How can I take the date in mm/dd/yy format and change it to YYYYMMDD format?

etc...

Sure, they can be handled in the respective forums, in the respective language, but they could certainly be discussed as an algorithm, too. I just don't think anyone would post to the algorithm forum and instead ask it in the specific language forum.

My $0.03

---John Holmes...

Ok, in the area of "can't find a way of doing that"; I was trying to write a piece of code which allowed for the selection and merging of cells in a table (like you can do in for example Dreamweaver), it's actually the first problem ever I've not been able to solve in at least a satisfactory way; whenever I fixed one problem, a old one resurfaced. It's close to a year ago now... never got around to doing it... The problem was of course taking care of all the rowspans and colspans and keeping the table sound. I got pretty close, but this was the sort of problem that needed to be solved completely or not at all.

(Oh if anyone has code for this sort of thing I would be *so* happy)