(Disclaimer: I'll freely admit that I'm interested in a language few people care about...I myself have a love/hate relationship with it. It's called Rebol, and was designed by the architect of the AmigaOS. Karmically, the company that has spent well over a decade designing it is a mirror of the Amiga contrast: technologically prescient, but chained to a community/integration vision that is a pile of fail. I digress...)

My issue is regarding this question that was closed by fairly anonymous moderators (as far as profile contact info goes)...calling it "not a real question":

Reading large Binary files fails in Rebol

Let me speak as one of the O(100) people who know this obscure language reasonably well. And I'll also go on record as saying this RebolTutorial fellow who posted it is neither a native English speaker, nor generally particularly lucid in his SO queries. But to those knowing the language, this PARTICULAR question is quite clear. (As a study in contrast, you could throw a dart at questions he's asked and likely hit one I'd myself call "not a real question".)

(UPDATE: Since I speak the language, let me rephrase what he said to make it clearer to those who do not...)

In the past, I've successfully used the read function to take a URL (or file specification) and load that information into a series data type. On small files it has always worked fine. But I'm trying to download quite a large file... the ISO of an Ubuntu disk image:

base-url: http://mirror.bytemark.co.uk/ubuntu-releases/lucid/
filename: %ubuntu-10.04-desktop-i386.iso
buffer: read/binary rejoin [base-url filename] 

When I try this it fails because of an out of memory error. Does anyone have sample code for how I might download such a large file, perhaps in chunks that do not require loading the whole ISO into memory at once?

I'm rather convinced that those closing this question have no clue about the language. So it would be like me going in and closing a question about perl or php, simply because I feel (as I do) that the languages are too unimaginably crappy to even exist. Yet I don't push that viewpoint, I just ignore those tags because they're not for me - "for the sort of person who likes that sort of thing, that's the sort of thing that sort of person would like..."

My leaning is to say I'm quite confident these "closers" couldn't write "Hello World" in Rebol. So should there be a CAPTCHA for certain tags...basic certification just to prove that you have the remotest qualification to deem something "a question" or "not a question"? Or are obscure languages doomed by the whim of people who don't know them?

  • 4
    Voted to reopen. However, it should really be edited to make it more "lucid".
    – user27414
    Jun 14, 2010 at 11:45
  • Sigh, yes. It would help if he'd put code in the proper indentation blocks and obey a convention or two. But frankly, within the Rebol community...using a communication medium that isn't built with Rebol is an ultra-rare flash of lucidity. Amiga die-hards are far more pathological than the Haiku OS folks, and I'm nutty enough to be drawn into these alternate realities. And in the Haiku case, Google even caught the fever... SEVEN summer of code students? haiku-os.org/news/… Jun 14, 2010 at 11:58
  • 4
    @Jon Do you honestly think this is a good question? If not, why reopen?
    – nb69307
    Jun 14, 2010 at 11:59
  • 5
    Following George's edit it now looks like a real question. Jun 14, 2010 at 12:12
  • 3
    @hostile, your edit is very strong. If you were to put that much care and effort into the Rebol questions that you see, many of them will remain open. And more "non-rebol" users would be inclined to read and get interested.
    – devinb
    Jun 14, 2010 at 12:28
  • 1
    Even if @hostile can't edit, by posting a comment to the question with a suggested edit change some editor would see it and edit the question to include it. Jun 14, 2010 at 12:32
  • 2
    @devinb: As a Wikipedia fanatic, editing would be pretty sweet. But I've programmed for a loooong time...so my modern interests are on the fringe. Though I try to make a good go at answering questions no one else will, it's not exactly the shortest path to reputation, especially when the folks posting obscure questions won't close them!! Sigh. stackoverflow.com/questions/2923647/… Jun 14, 2010 at 12:59
  • 1
    While we are complaining that users are not using conventions to quote their code, I would like to add a complaint too. this is rebol code read http://www.rebol.com but note that the http is removed from the visible comment on SO comments, though not here. So, following the convention destroys the syntax. Jun 22, 2010 at 23:22

6 Answers 6


As others have said, the main problem is that the person who posts most of the questions in the rebol tag isn't easy to understand, and (if my SQL serves me) of all the people that have ever posted an answer on a rebol-tagged question, only 11 have sufficient rep to edit them, so it's unlikely rebol questions will be fixed up by someone sufficiently knowledgeable to do so. The question really was just difficult to understand, and unfortunately it's unlikely that someone who both can edit it and understands what it should say will stumble across it


I think this question absolutely should be closed as "Not a real question".

That doesn't mean that the question (at its heart) is invalid, it means that the question as currently posed is invalid.

As you've explained in a comment

read/binary http://mirror.bytemark.co.uk/ubuntu-releases/lucid/… IS Rebol code.

Then the post says it "fails out of memory". Then the post says "Sample code?" At that point it is completely off the rails. He is asking for sample code to fix and out of memory exception? He is asking for sample code to do something he can already do?

The question is vague and the information provided is almost completely useless. Perhaps those who know Rebol can extract the question, but it is very much not-well-formed. For that reason, you should edit the question into coherence. It wouldn't get closed if the question, as formulated, made more sense.

If all the Rebol-relevant stuff was replaced with C, I would still vote to close, because the question is extremely unclear.

The fact that it is an obscure language certainly makes it harder for others to understand it, but the fact remains that if it were a strong question to being with, then it would not have been closed.

EDIT The question has now been edited into a much clearer and readable format thanks to George Stocker


I would have voted to close it too.

Here's why:

The question is but a sentence long, and it doesn't coherently describe the OP's problem.

Yes, it's a valid programming language, but with so little detail, it's impossible to ascertain what's going on. From the text:

Not a Real Question

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, or rhetorical and cannot be reasonably answered in its current form.

The OP also pointed to a .iso file. I don't really see any code there. If Rebol is some sort of binary only language and has no source code, then perhaps Stack Overflow isn't the place for it?

If Rebol is a language that exists in text format, then why didn't the OP post the code that he tried that didn't work?

After Editing The Question:

I also noticed that the OP failed to use backticks to indicate code (highlight and press ctrl+k); had he done that, it's possible people would have been more easily able to distinguish it as code.


Bottom line: You've got to put effort into your question if you don't want it closed.

  • You of course improved the question, no doubt. But I'll mention that Rebol does not require you to put quote marks around things like email addresses or URLs. The fact that you can't use : or @ in your variable names is enough to cue it to know those should be interpreted as "email literals" or "URL literals". While avoiding quote marks is not overwhelmingly compelling in and of itself, Rebol's fluidity has some deep semantic ideas which do tie it very strongly to how we use natural language. Jun 14, 2010 at 12:43
  • 1
    The 'quote marks' for the URL doesn't have anything to do with my answer. The Original Poster did not use the standard Stack Overflow code convention to indicate that he was posting code. (Which is ctrl+k). Had he done that, we wouldn't be here. Jun 14, 2010 at 12:54

I'd guess it was closed because the OP posted a link rather than actual code - I'm very anti that sort of thing myself. As you observe RebolTutorial is not only often not lucid, but also quite rude, which may have got peoples hackles up. But I don't think it was because the language is not used much - Smalltalk (similarly not used much) questions don't get that treatment, for example.

  • Well, you just don't realize that read/binary mirror.bytemark.co.uk/ubuntu-releases/lucid/… IS Rebol code. The way it tokenizes is wacky and fascinating; imagine an alternate universe that used LISP as its starting point instead of C. Yet because of the pathologies within Rebol community and lack of open source implementation, I'm hard pressed to suggest savvy C++ developers like yourself should look at it. Stare long enough into Rebol and Rebol will stare back into you. -- "Why, oh why, didn't I take the blue pill" -Cypher :) Jun 14, 2010 at 11:52
  • 2
    @Hostile It may be Rebol code, but as I don't normally follow links, I'll probably never know. If Rebol cannot be posted as ASCII text (from the Rebol site, it surely can) then SO may not be a good place to ask questions about it.
    – nb69307
    Jun 14, 2010 at 11:57
  • @Hostile - maybe formatting it as code would help to recognize it as code, especially for a not so popular language. I also agree that the question should not have been closed, but it is pretty hard to understand the real problem; it could have been a bit more explicit.
    – user141148
    Jun 14, 2010 at 12:04
  • I've updated my question to include what he might have written if he were more lucid. But I've seen worse stay open w/20 responses. My point is more on the topic of why folks are closing questions on languages they don't understand, when they don't comprehend the question, while sentient responders clearly do. Sunanda and I knew precisely what he was asking, and I posit that the people voting to close don't have an inkling about the language. It's like I said: if I had the power, should I close brief questions in languages I don't have a clue about because of spelling, indentation, etc? Jun 14, 2010 at 12:19
  • 1
    @Hostile You may have had a case if the OP had use standard conventions like code indentation when writing his question, but he didn't. In this case it would have been avoided if he had taken the time to properly format his question. Jun 14, 2010 at 12:21
  • @George: It certainly could be done better. But kid gloves are used on SO with questions like "how do I make a pointer to a class?" while there wasn't a single piece of feedback about the formatting or length, just an impersonal close notification (not easy to contest). I feel some frustration that popular languages with eager answerers seem to handle much newbier foreign noobs more gently than they treat noobs in languages they don't know. I'm pretty darn gentle, yet I caught flak for this! stackoverflow.com/questions/3025151/… Jun 14, 2010 at 12:32
  • Restating my premise, I might say: "When voting to close a question about a language one does not know, should the answers be considered in addition to the question...lest one be hastily judging the apparent clarity of something one knows jack about, when the relevant community can read it just fine." Jun 14, 2010 at 12:34
  • @Hostile But kid gloves are used on SO with questions like "how do I make a pointer to a class?" - poppycock - that would get closed pretty quickly on the C++ tags.
    – nb69307
    Jun 14, 2010 at 12:35
  • @Neil - 1. Poppycock is a real product you can buy in stores ( en.wikipedia.org/wiki/Poppycock ) and 2. questions far more n00b than this Rebol question get answers all the time instead of closed ( stackoverflow.com/questions/3033556/fifo-implementation ) ... nothing profound about that one, just saw it yesterday. Jun 14, 2010 at 12:47
  • @Hostile: That may not be the greatest question, however it is recognizable as a question. In fact, it's recognizable as four questions, which is a whole 'nother problem... In your case, an uncommon language combined with exceptionally lousy writing skills to obscure the question. FWIW, I refrained from closing it yesterday after seeing that you appeared to understand it in your answer, but I couldn't edit the question to forestall others closing it because I didn't understand what he was asking... Let's hope you get editing abilities soon!
    – Shog9
    Jun 14, 2010 at 14:49
  • @Neil: I answered a smalltalk question that you also made a note about. And I don't even bother with the language, I downloaded it because of the question. :) stackoverflow.com/questions/2081985/… Jun 15, 2010 at 7:58

The following is apparently valid rebol code:

read/binary http://mirror.bytemark.co.uk/ubuntu-releases/lucid/

is this next block valid rebol code?

read/binary http://mirror.bytemark.co.uk/ubuntu-releases/lucid/ubuntu-10.04-desktop-i386.iso fails because of out of memory.

That's what the original version of the question would have looked like if it had been indented. I'm betting that ending the line with "fails because of out of memory" makes it not be valid rebol code.

So, if what was posted isn't valid code even in the obscure language rebol, then how can you expect anyone to guess that this might be code at all? It looks like a badly-written sentence in English!

And, Hostile, if you had an issue with the question being closed, you didn't need to immediately run over to meta whining about moderators who don't understand rebol and "what to do" about them. What you needed to do is add a comment saying,

this is valid rebol code - or it would be if you ignore the part about "fails because of out of memory. It's a real question. The code means to read the file at that URL as binary. The OP is saying that when he does that for a large file, it fails with an "out of memory" error message.

The users who voted to close the question (they're not moderators, BTW) aren't at fault here for not knowing rebol. You are at fault for knowing rebol and for not speaking up!

  • To get needlessly pedantic (but we are all on the Internet because we like talking and learning, right)...it could be valid Rebol code. Although the OP didn't mean it that way. Y'see... Rebol is genuinely weird. If you loaded up the interpreter and had typed merely fails: func [bec of1 out of2 mem] [print "out of mem!"], and then because: 0 of: 0 out: 0 memory.: 0 then yes, that entire sentence would be a valid program. Make no mistake... Rebol is so insane that it could be cool, except that the people writing it are using Amigas and spend most of their waking life mocking Twitter. Jun 15, 2010 at 7:46

Wow I just came back from work: it seems I missed an epic battle here :)

OK I confess my english is generally wacky but as for this precise question, it was short enough just by reading the source code which is ONLY ONE LINE in Rebol whereas it would maybe be 100 lines in other languages.

If I had asked this question:

how to read a big binary file in C#

I guess everybody would have understood so why

how to read a big binary file in Rebol

would be less clear ?

The problem is that people thought that when I wrote

read/binary http://ubuntu.iso

it was some weird english whereas it was ACTUAL SOURCE CODE Rebol is just very concise, It's just ONE LINE :)

When some people ask questions about Ruby or Python they do formulate their questions to their peers not to people who do not understand Ruby or Python, so why would I formulate my question otherwise ? My question is for people who know Rebol, why would I target those who don't ?

Overall I'm satisfied as I got answer to my question so thank you all and above all to HostileFork !

  • 2
    where you went wrong, (as many have mentioned) is not using the constructs Stack Overflow provided for showing that a string of text is code (by using backticks ` around the code section. Had you done that and clarified the title, it would never have been an issue. Jun 14, 2010 at 19:01
  • Maybe it's again my english problem as I don't understand your remark about backticks: I did use stackoverflow icon to transform the line into code so I can't see why I did something wrong. Jun 14, 2010 at 19:09
  • 2
    "how to read a big binary file in Rebol" - that's not what you wrote. Not even close. It's almost what George wrote, which is probably what you should have written... And as George notes, it's not as though you marked your code as code, so SO tried to be helpful and interpreted part of it as an actual URL, leading to even more confusion.
    – Shog9
    Jun 14, 2010 at 19:11
  • I updated my answer with a picture. meta.stackexchange.com/questions/53547/… Jun 14, 2010 at 19:19

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .