Talk:HTTP

Clarify

Again, the task specification is really vague. It only says "Print a URL's content." This rises questions:

• What do you mean by "print"? Print on paper or display on screen? There seems to be both interpretations in current implementations.
• What do you mean by "content"? The source code or rendered page?

Why is it so difficult to write even couple of sentences to specify the task? Specifying the task is the most important part of any software project.
--PauliKL 09:30, 3 October 2008 (UTC)

Relax, it's fixed. --Mwn3d 12:32, 3 October 2008 (UTC)

I'm not sure I like the approach. Wouldn't "Display the request content using the most convenient method" be more appropriate? Some languages may not have easy access to the console. --Short Circuit 03:54, 4 October 2008 (UTC)

That's OK too. --Mwn3d 16:10, 4 October 2008 (UTC)

Fixed? I don't understand this spec. The content of a URL is the interior of the string. What is "source code"? Where in the HTTP specification do they use the term "source code" to denote any portion of the document named by a URI?

(Please sign your posts). Comments on the change in wording? --Paddy3118 00:54, 20 September 2011 (UTC)

Please confine your attention to what is said, not who. I see the text "Paddy3118, ...", but that means nothing to me and carries no authenticity (not that any is required: you're not writing me a check).

This is one example why you need to sign your posts. You inserted a paragraph in front of someone else's comment without a sig, at the same indent level, now it's difficult to tell if the above utterance was from Kernigh or some impolite person. If you are too good to register a user name, at least have the courtsey to sign it with your IP. --Ledrug 21:48, 20 September 2011 (UTC)

If you fetch an HTML file, you get HTML source code. This is not true in the general case; http://rosettacode.org/mw/title.png is not the source code of anything. --Kernigh 01:08, 20 September 2011 (UTC)

Yea, it locates a resource, the nature of the resource is carefully not defined. --Paddy3118 01:25, 20 September 2011 (UTC)

If you fetch an HTML file, you get an HTTP response with headers, and a body which contains HTML code. Some of the progrmas in this page look like they just dump the raw response, whereas others show the body.

Not Tested

Well..someone can test C# and Ruby examples? They are working?

ruby is working Rahul 18:47, 5 October 2008 (UTC)

C# (on Mono) is working --ShinTakezou 18:34, 27 January 2009 (UTC)

Recent Changes

(eg) https://rosettacode.org/mw/index.php?title=HTTP&type=revision&diff=315692&oldid=304000
Someone has just spent an awful lot of time updating this page.
I would like to think they did so with the best of intentions, however:
They did this anonymously, without any prior consultation.
(For instance, I might argue that rosettacode.org itself remains a better choice than w3.org, and probably wikipedia/google too.)
They changed every single example, almost certainly without proper testing for the majority.
They removed quite a few comments.
They removed all output sections.
There is a C example buried in there that might be worth preserving, as an alternative to the original (or v-v).
Two JavaScript examples for XMLHttpRequest() and jQuery were replaced by a single fetch() example.
Twenty "omit from" have been removed.
I am not in any way personally upset by this, but can see that if left it might set a dangerous precedence: discuss.
--Pete Lomax (talk) 13:30, 6 November 2020 (UTC)

It would be good if this user (or users -- I see several user ids following the same pattern here) spoke up and talked through these issues.

That said, looking at https://rosettacode.org/mw/index.php?title=HTTP&type=revision&diff=315692&oldid=304000 I agree that the missing output sections should be restored. (Possibly with some markup inserted if there's some crude cdn or anti-malware system we're offending somehow.) --Rdm (talk) 16:38, 6 November 2020 (UTC)

I agree with Pete Lomax - there is strong risk that a very dangerous precedent is being set. The solutions have been subsequently modified and probably they will work (not sure about the AWK version). The Batch File solution has been removed with little justification - are we going to let new users arbitrarily decide what is a programming language and what isn't or what version of a programming laguage should be used or what comments are "allowable" etc. ?

That all these massive changes by new users didn't receive any moderator comment until Pete Lomax pointed it out is also very concerning. --Tigerofdarkness (talk) 18:33, 6 November 2020 (UTC)

Similar stuff is now also occurring on Concurrent_computing by User_talk:PUtsApGpN7. Thundergnat appears to be aware of the situation and has already reversed changes to JSON (see User talk:NUKnWJiJTn) and posted a remark about comments to User talk:LVrOk6FqRY. --Pete Lomax (talk) 18:56, 6 November 2020 (UTC)

Thinking about this -- the output sections were removed for code where the url being used was changed. This suggests that those changes went untested. Which, itself, is a cause for concern. (I am debating with myself whether I should go through and revert the changes for those seven entries...) --Rdm (talk) 10:34, 7 November 2020 (UTC)

I am aware of the recent bout of mass changes to several tasks. I have several large projects I am trying to juggle for $work so haven't been able to dedicate as much time to my non-bill-paying activities. I am mostly ambivalent about changing the HTTP URL to w3.org, possibly mildly against as I'd prefer not to imply association or place uninvited load on their servers. I haven't really looked closely at the (140+!!) changes made to the HTTP task yet, though a quick glance shows several that are just wrong. My inclination is to just roll them all back since I really don't have time to review each one carefully. --Thundergnat (talk) 12:10, 7 November 2020 (UTC)

Sigh, too much time has passed to do a simple "Rollback all changes", I started doing piecemeal reverts but that got tiresome quickly and had multiple edit conflicts to resolve so I just ended up reverting to the revision current as of November 1st. There may have been some useful edits in there but they were lost in the noise. For future reference, from here on, if a new member with no history comes in and starts making large changes without discussion or consultation, I will summarily revert the changes and ban them. I don't have time or patience for this. This time it will be one week as there wasn't a policy in place. In the future it may be permanent. --Thundergnat (talk) 12:45, 7 November 2020 (UTC)

Thanks for your efforts, Thundergnat. I realise that being an RC moderator is not something that you get paid for. --Tigerofdarkness (talk) 15:06, 7 November 2020 (UTC)

RE: Recent changes

Hi everyone I'm the one who made all those changes.

The reason for the name is because honestly I disagree with the requirement for an email, password, username to edit on this website, and so created a random username, random password, and used a temporary email to make this account and all others I used to edit pages.

The reason for the edits, on the HTTP task specifically, is because the task was not specific enough, allowing solutions to be accepted if they redirected from HTTP to HTTPS, or even if the URL lead to a HTTP response code of Permanently Moved.

I specified the task so that a standard result was expected from solutions (plain text HTTP response shown on a screen, you seriously can't get more simple yet specific than that). I edited all the solutions to change ONLY the white space, the URL that was used (so that the URL was standard in all solutions, which seems to me to be such an obvious requirement that I doubt how dedicated to this webpage the person who created the original task is), and then removed the output. I removed the output because, since my task was so precise in its requirement, it was completely clear that all solutions would output the same response anyway so why include output?

You've already removed all the changes, which is honestly fair enough.

However I have a serious problem with you removing some requirements for the task which seem as obvious as requiring a URL to fetch in the task.

It's obvious that in the ENTIRETY of this website it should be a requirement for solutions to use external (non standard) libraries ONLY if the task is PHYSICALLY IMPOSSIBLE to solve without the external library.

It's honestly baffling that whoever is behind this website didn't think of this, and I've lost the motivation to help develop its content.

Cya, if I could delete this account, then I would, however it doesn't even look like I can do that, so I'm just going to forget it and let it die like all the other ones.