Esolang talk:Community portal

From Esolang
Jump to: navigation, search

Archives: June 2005‎ – February 2012‎

Site move / wiki read-only


Sorry about the spam everyone! I haven't been the most helpful site admin lately. Partly, the problem is I don't have a lot of free time to install better spam protection, and partly, my server is just kind of a mess (and it's shared hosting so I can't fix it).

But, good news! User:Ehird has offered to take over hosting the wiki. He promises to continue producing database dumps, keep the wiki content public domain, and install much better spam protection.

I'm about to make the wiki read-only so the move can commence. When it's done, the URL will point to the new server and everything will just automagically work (one hopes). As for, that will still point to my server where I'll continue to host the esoteric files archive and have a link/redirect to the wiki at --Graue 09:15, 17 February 2012 (UTC)

Ascension address

Hello! If you're reading this, you're on the shiny new Esolang. As explained above, after the wiki became overrun with spam, I contacted Graue about taking over hosting of the wiki. That has now happened, and within a few days everything should point properly at this new host.

The site is now running on the latest version of MediaWiki. All the important data has been transferred over, although watchlists and blocks have been wiped. Administrator rights have carried over; I've given myself and User:ais523 bureaucrat rights, and installed a range of spam-fighting and administration extensions, along with a new CAPTCHA. (For administrators: AbuseFilter and SpamBlacklist are the ones you can control.)

As a result, the spam problem should now be over; unless the spambots read up on their esoteric programming language trivia, the only way they can spam will be by reusing already-existing accounts they have. I'm going to look into deleting their accounts, but until then, blocking them on sight should suffice.

Additionally, I've changed the public domain dedication from the deprecated Creative Commons Public Domain Dedication and Certification to the CC0 1.0 Universal Public Domain Dedication, which is designed for international use and handles situations where releasing content into the public domain isn't possible (for example, due to the copyright laws of certain countries).

Graue will continue to maintain the file archive on his server, and I plan to restore public backups of the wiki in the coming few days.

Some additional improvements:

  • <div> and <span> tags now work without any special template hacks.
  • Certain previously-problematic page names, such as Subleq+ and ///, now work properly.
  • I've set up a lot of caching, so the site should go a lot faster in general.
  • Tables can now use the wikitable class.

If you have any technical problems, please contact me, or leave a message here.

Happy (verb form of "esolang")ing!

ehird 04:36, 18 February 2012 (UTC)

It's great to see the wiki back in shape and running smoothly. Thanks! --Keymaker 11:46, 21 February 2012 (UTC)
Would you ever install my TeX format file onto this wiki? I mean, the one designed for secure remote access. And it works better than all the other ones the other ones don't work --Zzo38 20:27, 22 February 2012 (UTC)
Sure, if you have a MediaWiki extension, I'll take a look at it. —ehird 07:04, 24 February 2012 (UTC)
OK, one of the spambots just somehow managed to pass the CAPTCHA. I'll look into this later today. —ehird 10:54, 24 February 2012 (UTC)
It seems like the CAPTCHA was solved by a human; the requests were slow and looked like a typical browser's requests, including downloading all the external resources like the scripts and images, and the IRP page was loaded after the registration page. The user agent looked fairly standard for IE 8, so I guess the only thing that can be done if this gets more common is using the filters to block the spam based on its content, and blocking IP ranges used by spammers. Hopefully, that won't be necessary. —ehird 13:06, 24 February 2012 (UTC)
Selecting from the "Special pages" dropdown menu doesn't work. (Did you forget to install redirect.php file?) --Zzo38 19:21, 24 February 2012 (UTC)
Fixed; redirect.php was installed, just not exposed to the web. —ehird 19:30, 24 February 2012 (UTC)

SpamBlacklist disabled

I've disabled the SpamBlacklist extension, since it's blocking too much (Oerjan couldn't save a page with a link to before I disabled it, even though the page already had a link to the same site). The Wikimedia Meta-Wiki blacklist the rule came from is probably a bit too broad and not that helpful for our spam, since all our spam links tend to be from a few sites. As our local blacklist is currently commented-out, we don't really have much use for it. Just a heads up for any admins that might otherwise wonder why adding entries to the blacklist doesn't do anything. —ehird 21:21, 6 April 2012 (UTC)

MediaWiki 1.19 upgrade and DNS change

I've upgraded the wiki to MediaWiki 1.19. has a summary of the changes. Probably the most prominent is that the default signature now includes a talk page link, which I will demonstrate at the end of this message.

Additionally, the DNS for has moved from to Linode, the VPS provider for the wiki. has had some outages recently, so hopefully this should be more reliable. This also means that you can now access the wiki via IPv6. Which is good, because the future is forever. We're even a month early!

As always, let me know if you run into any problems. ehird (talk) 21:48, 3 May 2012 (UTC)

Can reliability be traded for low power?

If a computer processor or volatile ram-memory gets barely enough electricity to work, is there a range of power where they just make mistakes relatively often, like one in thousand or one in million bits get physically changed by physical errors, or is there sharp threshold near currently used power levels where they just stop working if power gets 0,1% lower? I guess this question has to be asked separately for every different kind of semiconductor manufacturing method, clock speed and working temperature, but any knowledge about currently used common hardware?

If it is possible to trade-off low level reliability for power use that is half of normal levels, then it would be usefull to have a special operating system that can usually function in such computer long enough. Long serial computation would be less suitable than big parallel computation in search of an answer.

In that kind of trade-off, ideally, a computer would have one small reliable processor and many cores in a low reliability high speed per watt processor. That may not be practical for small budget, so maybe tiny reliable computer on firewire port for direct ram access?

Talk:Esoteric Operating System

Tektur (talk) 05:42, 22 May 2012 (UTC)

Maybe this was what inspired you, but apparently yes. --Ørjan (talk) 20:19, 22 May 2012 (UTC)

The lang at mailing list

I recently subscribed to the mailing list that this website calls "dead." The lang at mailing list. Well, some people still like mailing lists, so here's an idea. Let a bunch of us Esolang wiki readers subscribe to the mailing list, and start posting questions and news that is actually esoteric programming-related. Pump some life into that thing! Let those tape loops keep on turnin'! Star651 (talk) 21:19, 5 June 2012 (UTC)

The list is dead; the rudimentary archiving has been broken since 2009, and almost everything in that archive from 2008 onwards is spam. The mailing list software is unmaintained and barely working, too; it's not actually possible to subscribe to some of the other lists on that server any more. I also asked the list administrator (atehwa) on IRC, and he told me he didn't have the time to administrate the list (although he did say he was willing to delegate to others if it needed to be done, I suspect the other factors would make this difficult). So I wouldn't recommend trying to revive that list.
That isn't to say that a new esoteric mailing list couldn't be made active, of course, but I suspect it'd have difficulty gaining momentum, since the wiki is probably more convenient for discussion for most people. ehird (talk) 14:30, 7 June 2012 (UTC)
As I said on the email list, I don't know about others, but for me, email is more convenient by a long shot.
If the issues you point out are important enough maybe the list should move to a Google Group or something (at least if there's interest). -- 01:49, 8 June 2012 (UTC)

WPU and esoteric processors

By recently viewing the few images that this wiki has here, I noticed this "Weird Processing Unit" thing here: AttoASM and realized it relates to this and this. How these pages and concepts should be fitted together, allocated points of views etc.?

Tektur (talk) 02:37, 9 June 2012 (UTC)

Names for ideas

Please give some name canditates for the ideas in list of ideas page. I think the names should be in english, and words or acronyms, something you can pronounce and type to a search engine, and also not conflict with other strings related to programming (or preferably anything).

Esoteric sub-operating system, sub-OS

There has been some pondering to whether or not an esoteric operating system should be build, and what that could be. What about an OS that is meant to be emulated like windows is emulated in linux by wine? Then that OS or actually sub-OS could run as a linux, windows or mac application and run sub-applications / plug-ins that run on any actual operating system that has eso-sub-OS for it.

Tektur (talk) 06:36, 26 June 2012 (UTC)

Depending on what it is, it could be called auxiliary operating system.

Tektur (talk) 18:33, 1 July 2012 (UTC)


As I suggested in #esoteric earlier, I would like to have "infoboxes" on language-related pages of this wiki. I think that they will make it easier to find information quickly and will make sure that authors give basic information about their language (this doesn't always happen). In addition, I also see it as a form of organization. This basic information would include:

  • Name
  • Logo (if available) with caption
  • Year of creation
  • Supported paradigms
  • Type system
  • Designer / Inverter
  • Reference implementation
  • Influenced by
  • Equivalent languages (maybe dialects?)
  • File extension(s)

Anyway, let me know what your opinions are on this idea. --AnotherTest (talk) 09:05, 14 October 2012 (UTC)

I think all those items are good except type system - I cannot at the spot remember any esolang which I am sure has a static type system, although this may be because I tend to ignore esolangs that look like nothing more than a mixed heap of features from everywhere. Some may have one as a side effect of being implemented by translation to something like C++. In any case it is not usually a prominent feature. -- Ørjan (talk) 10:01, 14 October 2012 (UTC)
You are right, the type system should probably not be included. Maybe a link to a specification would also be nice? Although I don't think many esolangs have an official one. --AnotherTest (talk) 10:22, 14 October 2012 (UTC)
"Equivalent" languages seem too restrictive - something symmetric to "influenced by" would be nice, I think. Most 'official' specifications are either the wiki page itself, or linked to under external resources, I guess. You could also add computational class and stuff like that, but then it's starting to be redundant with categories... --Koen (talk) 18:24, 14 October 2012 (UTC)
So I looked up how to technically realize this, a duckduckgo search brought up this webpage, which seems to be a rather nice explanation. --AnotherTest (talk) 18:55, 14 October 2012 (UTC)
Please don't follow that guide on this wiki; Wikipedia's templates are licensed under an incompatible (overly restrictive) license compared to Esolang's CC0 public domain dedication, and hence it would constitute a copyright violation. The templates would have to be deleted. ehird (talk) 19:42, 14 October 2012 (UTC)

I created a template Template:Infobox proglang. GermanyBoy (talk) 09:14, 5 July 2014 (UTC)

List of "good" languages

As I suggested earlier on #esoteric, I believe we need a list of languages that can be presented to new visitors as "good esoteric languages". Let this criteria be well specified, so that a language may be removed from the list objectively. For this list I would set the following requirement (please extend this list with your own ideas):

  1. Personally I think that the languages on the "good language list" should be implemented. I realize that some very good languages have not yet been implemented, although visitors should be able to try out these "good languages".
  2. The language should have a decent specification (either external or on the wiki); this means that it is correct both in language/presentation and in content. It should also be complete.
  3. The language should not be under development, it should be more than a concept.
  4. The language may not be convertible to another language by replacing lexemes from one language with lexemes from the other language. Note, this applies only when there is a "one by one" replacement.
  5. The language must contain a new or unique concept. This is of course hard the objectively determine, but let us assume that making a brainfuck version with a tape of complex numbers is not unique enough.
  6. It is preferred (but not absolutely required) if the languages are of historical importance to the esoteric programming language community (eg. brainfuck).
  7. Something notable (eg. an interpreter, a prime sieve, a quine that is not the empty set of instructions...) has been implemented in the language. This does not include the "Hello, world" program, nor does it include a square root finder in a language that finds has a built-in operator to do this. This highly depends on the language. For example, a cat may be hard to write in one language, whereas it might be very trivial in another.

You are free to add a language to the list if and only if the language meets this requirements and it is not your own language. --AnotherTest (talk) 14:51, 17 October 2012 (UTC)

We already have Esolang:Featured_languages/Candidates. How is this proposal significantly different? Chris Pressey (talk) 08:21, 18 October 2012 (UTC)
The featured languages have to go through a voting process, and only the wiki administrators can add them. The result is it being a list of just 2 languages. --AnotherTest (talk) 08:06, 21 October 2012 (UTC)
True, re the process, but otherwise, the criteria are almost the same. If this proposal moves forward, it would seem to make the most sense to me to have the featured languages be drawn from this category (which could be called "Featured Language Candidates"), or something.
(Also, I think combining "can be removed from the list objectively" with several of these criteria is just asking for trouble. e.g. "preferred if of historical importance" doesn't let me objectively decide if I can remove a language, because it's just a preference.) Chris Pressey (talk) 13:44, 21 October 2012 (UTC)
A rough simulation of the criteria on this list with a single objective criterion might be: are there two (or more) implementations of the esolang, by different authors, in different host languages? The idea here is that the existence of more than one implementation implies that the language is developed, interesting to someone besides the designer (interesting enough to motivate them implementing it, in fact) and sufficiently well-defined that it was able to be re-implemented (which suggests the existence of a specification and some test cases.) I raise this possibility because it actually is an objective criterion, while any guidelines which include subjective criteria for "good" will probably be bogged down somewhat with review/consensus/editwar processes. Chris Pressey (talk) 14:31, 29 October 2012 (UTC)
That would rule out Banana Scheme, TwoDucks, Gravity or You are Reading the Name of this Esolang, however. Unless you consider that an implementation of, for instance, Brainhype in Scheme-omega counts - but it would certainly not allow for test cases... --Koen (talk) 17:34, 29 October 2012 (UTC)
I realize that, and probably should have mentioned it as one of the limitations that makes it a rough simulation. For unimplementable languages, maybe a roughly equivalent objective criterion would be that two different people wrote programs in it accompanied by proofs of their correctness. (Which would still exclude some languages, but you can't hope to get them all with this sort of approach.) (btw, iirc, You are Reading the Name of this Esolang can actually be implemented -- the spec only requires that an implementation uses a technique that finds some always-halting programs and some never-halting programs. It doesn't insist that an implementation detect them all.) Chris Pressey (talk) 22:03, 29 October 2012 (UTC)
Actually so far the process is "Anyone nominates, ehird chooses the next candidate from the list", and I think the main reason the featured languages are currently stalled is that he thinks all of the better current suggestions need editing improvement before promotion. Otherwise, we are way overdue for a new one. (And there isn't much point in a voting process unless there is more than one article already of sufficient quality.) --Ørjan (talk) 14:30, 21 October 2012 (UTC)
I think your own language should be allowed too. In addition, descriptions should be required, and should require multiple nominations, too. Also, it should be sufficiently old. If it is too new then maybe you missed something. --Zzo38 (talk) 19:19, 29 October 2012 (UTC)
If the criteria are truly objective I see no reason to prevent anyone from adding their own languages; there would be no possibility for bias. Chris Pressey (talk) 22:03, 29 October 2012 (UTC)
Sounds like all we need is just a more open voting (or other decision-making) process. Why should it be limited only to admins? — Timwi (talk) 11:48, 27 January 2013 (UTC)


Looks like should be used for esolang related open source software more than it is used now. And this wiki needs a page github that explains it from esolang point of view. Maybe it should be mostly text from this: wikipedia:github ? --(this comment by Tektur at 14:31, 4 November 2012‎ UTC; please sign your comments with ~~~~)

Does Github really need an explanation from an esolang point of view? A macro to link to a Github repo-page might be useful, but beyond that, Github's just another of the many places that esolang descriptions/implementations can be hosted. (Also, full disclosure: essentially all of my esolangs are in repos on Github.) (Also also: we can't copy text out of wikipedia into this wiki at any rate, for licensing reasons.) Chris Pressey (talk) 16:25, 4 November 2012 (UTC)
I guess many other sites meant for source code could have their own pages too in this wiki. Those pages could be very short, 2 sentences and some links to esolang related parts on it. If so, what should be the categories for those articles?Tektur (talk) 10:50, 8 November 2012 (UTC)
I've been thinking idly about writing an article about "good" practices to follow when designing an esolang (describe it as clearly as you can; write lots of example programs in it; implement it; host the implementation in source form somewhere; put an open-source license on it or make it public domain -- in roughly that order of importance) and such a list of hosting options could be part of that article. But the "good" in those practices is probably somewhat subjective, so, I'm not sure where/if such an article would fit in this wiki. Chris Pressey (talk) 16:04, 8 November 2012 (UTC)

Crystal co-processor

It seems possible that some special crystal formed from carefully designed molecules could be an implementation of some 2d or 3d esolang that may or may not be already invented. Interactions between those molecules would mean transfer of bits. The molecules would be arranged in precise homogenous repeating form, possibly in cubic form with 6 neighbors in left, right, front, back, up and down, or some other arrangement. Even though such crystal computer or crystal co-processor would be computationally very inefficient, it's physical efficiency might outweigh that, at least in some kind of computation. Crystals and molecule design are routine parts of chemistry, and that would not be the most difficult part in making them. Input and output to that crystal would be, because it would need much smaller parts than current consumer level semiconductors have. But chemists and physicists are working on electronics that small, and already can do something at least in lab. If it needs extreme coldness with liquid helium or liquid hydrogen cooling, then it could be practical at least with big server farms / cloud computation services.

What is the closest esoteric language to that needed for such crystal co-processor? It may or may not be 2 or 3 dimensional or based on dimensions at all, even though it's derivative for the crystal would have to be.

In case somebody thinks so, I am not in a position to make anything like that crystal. Tektur (talk) 10:50, 8 November 2012 (UTC)

Sounds to me like some kind of asynchronous cellular automaton would probably be the best match. This is under the assumption that every molecule inside the crystal is in continuous and independent "operation" -- i.e., there is no central "clock", there is no single "instruction pointer", it's naturally "distributed", etc. I don't know of any esolangs like that offhand. Although there are a few cellular automata designs, and a few asynchronous designs, and maybe some of these could be combined. (I'm guessing it's probably easy/cheap to grow crystals with random computational properties... the difficult/expensive part would be getting ones with the computational properties that you actually want :) Chris Pressey (talk) 15:18, 8 November 2012 (UTC)
I think the crystal could be synchronized somehow, maybe with a radiowave / microwave or if it is transparent enough, with flashing laser. With microwave, it's phase would be the clock. By the way, normal microwave owen happens to be ROUGHLY same frequency as fastest electronics have been and probably will be: 2,4 gigahertz (when world record for a processor was 6-8 gigahertz) (not to say that it is the correct frequency for the crystal even roughly, it could be 100 gigahertz or 100 megahertz... (100Ghz and higher are used in radars and microwave communication links)).Tektur (talk) 20:19, 8 November 2012 (UTC)

One molecule could have thousand atoms if that kind of complexity is needed to implement a cell in that cellular automata. That means, we don't have to know much about chemistry in order to develop the rules and basic patterns for the cellular automata in a crystal. If anything like that is possible to do, any rules can be implemented and others are much better suited to develop the molecules.

I have read that even conway's life can be made turing complete with complex, convoluted and space wasting patterns and methods. The actual cellular automata used in crystals would be much more efficient. One type of computation crystal could have a normal integrated circuit putting input patterns on it's front side and a reader circuit on it's backside getting the results it outputs. On every clock cycle, every 2D cellular automata state gets computed one step, while physically moving to a layer one molecule width closer to the backside. If the crystal has million layers of molecules from front to back, it can hold and calculate million separate 2D cellular automatas at one time. Then every cellular automata will have exactly million steps / iterations. (million is just an example number, it could be 100 or 1000) That would make it a stream processor like graphics processors are. I guess true 3D cellular automata with real turing completeness is possible too.

If needed: -Crystal layer could have 3 or 6 link topology with adjacent molecules in the same layer, instead of 4 or 8. (with hexagons) (If we talk about chemistry, carbon atoms in a graphite layer have 3 links within the layer.)

-Rules can be different to different directions, for example different to front and back and same for 4 other sides. Tektur (talk) 22:20, 14 November 2012 (UTC)

I created a page for this topic: Computing crystal. Further discussion on it or it's talk page / discussion page. Tektur (talk) 09:42, 25 November 2012 (UTC)

Maybe esoteric programming is an adult/mature-content subject ... ?

No esolang website I know of has a warning. Now, don't misread this; I like esolangs, but then again, I have a Monty Python sense of humor. I'm just stating a fact. In a genre of computing/literature/programming where language names like "Brainfuck" and "Minifuck" are prevalent, wouldn't you call esolang an R-rated genre? Why not put a warning on the site? Star651 (talk) 20:56, 17 January 2013 (UTC)

Personally I don't really care about the opinions of the kind of people who would say the word "fuck" is enough to make a site adult/mature-content. --Ørjan (talk) 04:46, 18 January 2013 (UTC)


We need a template or category that we can put on a page to signal to admins that they can delete said page. Is there such a template/category already? — Timwi (talk) 11:35, 27 January 2013 (UTC)

Weird that nobody responded to this. If there is no such template or category, what is the accepted practice? — Timwi (talk) 14:18, 8 January 2016 (UTC)
Most admins patrol recent changes, so pretty much any sensible method of requesting a deletion will work (assuming that the admin agrees with you). Note that in the case of pages with useful history, it can make more sense to redirect them (assuming a reasonable redirect target) so that non-admins can continue to view the history afterwards. --ais523 21:51, 8 January 2016 (UTC)




The most tightly compressed Linux

If we have a source code compression format that strips all comments, replaces all names with codes and puts all outputtable text on a big string so zip can compress it better, and a minimal compiler and gcc compiler that are converted to compile that format directly, I think that would enable some Ubuntu level Linux to be distributed in a record small packet.

When it begins to install, it has minimal system with minimal non-optimizing compiler that as it's first and only action compiles gcc source code. Then the unoptimized but optimizing gcc compiles it's own source to get a faster optimized version of itself. Then that gcc compiles the rest of the system.

Minimal compiler could be replaced with interpreter for c and other languages if that is smaller. Then it would interpret gcc source code that has gcc source code as it's input.

All this can be automated.

One kind of version would first compile the installation software to ram memory and from there allow to choose what to compile and install, how to set hard disks etc.

Compiling takes time, so it is more useful with slower connections.

Tektur (talk) 20:02, 7 July 2013 (UTC)

Site move, redux

I've been neglecting the wiki lately; the MediaWiki installation is many versions behind and I've haven't really touched it in many months. I don't want to let it become dormant, so I'm handing administration of the wiki over to fizzie (not terribly active on the wiki himself, but a long-time community member and operator in #esoteric), with hosting provided by GregorR. The wiki will be read-only for a few days until the transfer is over and the DNS change propagates; sorry for the inconvenience, and thanks for flying Esolang Airlines! ehird (talk) 16:26, 23 March 2014 (UTC)

If you're reading this, then you're on the new server. My connection picked up the DNS change sooner than I expected. ehird (talk) 19:04, 23 March 2014 (EDT)

New filter

Looking at the number of legitimate anonymous edits caught in the new filter (which admittedly has caught much more spam (and all of it) so far), and the fact that exactly one user has so far gone on to complete the new process, I am starting to think it's a little too brutal. And also potentially insulting to editors who aren't told they need to do any of this until after they've already clicked submit on their edit. --Ørjan (talk) 18:16, 31 August 2016 (UTC)