Limit and advisability to delete an answer

I recently incur in this question ("Can EXE always be replaced with COM?") and in the deleted answer that I report below for the users that have no access to the deleted questions.

It is an upvoted answer (~70), with many version (5) made by the author, with more than 15 comments. It must have been popular in the summer vacation period :-).

I have no problem to recognise some inspiration as proposed by other users, but I can see even some work in cutting the not essential parts (in relation to the question), a general paraphrase of the other page, some personal additions (in Z-80 / 8088 code).

We usually deprecate to post one-link answer asking to report the main points: here it missed the link (now there is one), but I cannot be sure that the author takes really that page as inspiration, or another page on internet (in principle even older) as starting point.

Nonetheless there is some personal work that differentiates the original blog page with the answer, and it is possible that someone finds it more clear and readable than the accepted one.

Note that the accepted answer was posted some hours after the deletion of this one, with the explicit link to the blog page, a similar paraphrase that simplify even more the concepts expressed and with the addition of a link to a stackoverflow similar question.

So I'm asking myself and to you all:

  1. is it that enough to mark as plagiarised and delete this answer (and others like this) ?
  2. is it not a damage for the normal user that the answer is deleted and not any more accessible (even more in consideration with the fact that it was really highly upvoted and so many people just thought it was useful)? In general the information inside the deleted answer should be get lost...

  3. is it not the case to undelete it, maybe with a critical note that states the absence of the source link, and let the people read and decide?

Thanks for your time.

The deleted answer in its last form

[paraphrased from Raymond Chen, Microsoft - What’s the difference between the COM and EXE extensions?]

In the beginning of MS-DOS (and CP/M) the only programs that existed were COM files.

A COM file was just a memory image (in Z-80 / 8088 code). To load a COM file, the loader used to put the file into memory and started it at the first byte.

The COM file format had the problem that programs were limited to 64KB. To manage that limitation the EXE format was introduced. The EXE file has a header that begins with the letters "MZ" and includes information that the loader uses to load the program into memory. COM files were "raw memory images" and EXE files were more "structured".

Over time the original MS-DOS programs like FORMAT.COM, EDIT.COM and COMMAND.COM grew larger than 64KB and they had to be changed to EXE, but it meant a compatibility problem with a lot of software produced in the past. It was solved by giving the loader the flexibility to check if the file begins with "MZ", independently of the extension, so as to know if it is a "structured" or "raw" file.

So nowadays it doesn't matter what the extension is (COM or EXE) because the loader will recognize the header and will manage the file accordingly.

answered Aug 11 '16 at 13:26 jcbermu

  • I don't see what is wrong, with the current current version of the answer, proper citation was provided. As for the current accepted answer, I have, more problems with it, then the one that was deleted.
    – Ramhound
    Commented Feb 27, 2017 at 18:53
  • That deleted answer looks fine to me as well. Unfortunately deleted by a mod so I can't vote to undelete :/
    – DavidPostill Mod
    Commented Feb 28, 2017 at 9:10
  • @Ramhound I'm too... I'm waiting for some answers able to explain different points of view, pro or contra... Not only in this particular case but in the general too. With the deletion was wasted the time used to research, write, correct, read, vote... and I'm not still sure if it was possible at all to say that it was plagiarism (even without the link), after that the OP reworked (and integrated) the original page.
    – Hastur
    Commented Feb 28, 2017 at 11:40
  • @DavidPostill Let's see if it will arrive some answers to shed light on... then I think we can flag it for moderators attention with the request to un-delete.
    – Hastur
    Commented Feb 28, 2017 at 11:42
  • @DavidPostill:  Guess what?   You can do something about this now.    :-)    ⁠ Commented Apr 17, 2018 at 2:26
  • @Scott Wake me up if something happens... he will have my vote for reopening -- if needed. (we should take in due consideration the time spent to do some work on the site...).
    – Hastur
    Commented Apr 17, 2018 at 5:37

1 Answer 1


I think that there's a real need to avoid plagiarism in answers, anywhere on SE. I also think that there is no need to reinvent the wheel. If the web has the answer, and there's little that can be added to that answer, then don't try. Providing the source of the information, preferably with a link, cancels plagiarism. Even the Help Center recognizes that answers will need external material. It does say to only use what's needed and to use your own words.

OTOH, the guidelines for good answers tells us not to provide links without context. The link could become unavailable, or even be deleted. So we need to include the relevant info from that source in the answer, so that the answer here can stand on it's own, if need be.

The older a technology is, in this case decades, the more likely it is that what the web has will be as terse as possible already. That leaves very little room for people here to summarize that info. (How do you summarize the overview of a summary?) For the question at issue, the reply could have been "An EXE renamed to a COM will not run in anything earlier than MS-DOS 2.11." Ok, I may have the wrong version of DOS, maybe it's 3.21, I don't remember. Nevertheless, that literally answers the original question.

To provide more info, and make the answer worthwhile, the deleted answer here provided more background, gathered from the web, and reduced it reasonably. In the form presented here, the answer includes proper source citation. Comparing the answer here to the cited material I'd say a credible job of condensation was done. Maybe there could have been more background, maybe less was needed to answer the question. Some "wiggle room" needs to be there to avoid simple answers that "work" for the OP but do nothing to make SE better. I suppose the answer could have added other material, not found in the citation, such as the "magic letters MZ are for the creator of the concept, Mark Zbikowski," but that's probably too much trivia for the answer anyway. It was a good, non-plagiarized, answer, and would have gotten my up-vote.

My opinion, flag it and get it back. It's better than what's there now.

  • Thanks for your time, you bring as example the same points (dos version and MZ) that I would like to add to that answer :-). [I should be worried if I were you ;-)... ] ps> there is another point for me important, time. We all present our time to our community, unnecessary deleting it is a waste of this time, so a pity.
    – Hastur
    Commented Mar 1, 2017 at 6:06
  • @Hastur I've nothing to worry about. I got what MZ means before you did. Don't quote me, though. BTW, not a legend, either. It used to be explicitly stated in the documentation about the EXE header. Not everything on Wikipedia is questionable. Just 98% :)
    – Chindraba
    Commented Mar 1, 2017 at 6:12

You must log in to answer this question.

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