Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #34976 > unrolled thread

Disappearing API docs

Back to article view | Back to comp.lang.java.programmer

Contents

  Disappearing API docs Roedy Green <inquiry@mindprod.com> - 2016-07-19 06:46 -0700
    Re: Disappearing API docs Eric Douglas <e.d.programmer@gmail.com> - 2016-07-19 06:54 -0700
      Re: Disappearing API docs Lew <lewbloch@gmail.com> - 2016-07-19 10:36 -0700
        Re: Disappearing API docs Roedy Green <inquiry@mindprod.com> - 2016-07-19 12:23 -0700
          Re: Disappearing API docs Eric Douglas <e.d.programmer@gmail.com> - 2016-07-19 12:49 -0700
            Re: Disappearing API docs Martin Gregorie <martin@address-in-sig.invalid> - 2016-07-19 20:31 +0000
              Re: Disappearing API docs Eric Douglas <e.d.programmer@gmail.com> - 2016-07-19 13:59 -0700
                Re: Disappearing API docs Lew <lewbloch@gmail.com> - 2016-07-19 14:52 -0700
                Re: Disappearing API docs Martin Gregorie <martin@address-in-sig.invalid> - 2016-07-19 22:03 +0000
                  Re: Disappearing API docs Eric Douglas <e.d.programmer@gmail.com> - 2016-07-20 05:36 -0700
                    Re: Disappearing API docs Nigel Wade <nmw@local.domain> - 2016-07-20 17:10 +0100
                      Re: Disappearing API docs Eric Douglas <e.d.programmer@gmail.com> - 2016-07-20 10:39 -0700
                        Re: Disappearing API docs Eric Sosman <esosman@comcast-dot-net.invalid> - 2016-07-20 13:59 -0400
                          Re: Disappearing API docs Eric Douglas <e.d.programmer@gmail.com> - 2016-07-20 11:08 -0700
                            Re: Disappearing API docs Martin Gregorie <martin@address-in-sig.invalid> - 2016-07-20 21:55 +0000
                            Re: Disappearing API docs Eric Sosman <esosman@comcast-dot-net.invalid> - 2016-07-20 18:19 -0400
                          Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-22 13:31 -0400
                        Re: Disappearing API docs Nigel Wade <nmw@local.domain> - 2016-07-21 10:08 +0100
                        Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-22 13:30 -0400
                    Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-22 13:16 -0400
                      Re: Disappearing API docs Gene Wirchenko <genew@telus.net> - 2016-07-22 15:24 -0700
                        Re: Disappearing API docs Lew <lewbloch@gmail.com> - 2016-07-22 15:34 -0700
                        Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-22 19:29 -0400
                          Re: Disappearing API docs Gene Wirchenko <genew@telus.net> - 2016-07-23 21:45 -0700
                            Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-30 22:33 -0400
                              Re: Disappearing API docs Gene Wirchenko <genew@telus.net> - 2016-08-02 11:07 -0700
                                Re: Disappearing API docs Eric Sosman <esosman@comcast-dot-net.invalid> - 2016-08-02 14:26 -0400
                                  Re: Disappearing API docs Gene Wirchenko <genew@telus.net> - 2016-08-02 15:53 -0700
                                Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-08-03 21:55 -0400
                Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-19 20:43 -0400
            Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-19 20:40 -0400
              Re: Disappearing API docs Martin Gregorie <martin@address-in-sig.invalid> - 2016-07-20 11:45 +0000
                Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-20 08:47 -0400
                  Re: Disappearing API docs Martin Gregorie <martin@address-in-sig.invalid> - 2016-07-20 18:44 +0000
    Re: Disappearing API docs Arne Vajhøj <arne@vajhoej.dk> - 2016-07-19 20:24 -0400

Page 1 of 2  [1] 2  Next page →

#34976 — Disappearing API docs

Oracle has renamed the files that form the Java API Java doc.

https://docs.oracle.com/Javase/8/docs/api/java/util/zip/ZipEntry.html
is now 
https://docs.oracle.com/javase/8/docs/api/java/util/zip/ZipEntry.html

Oracle's server is case-sensitive.

[toc] | [next] | [standalone]

#34977

On Tuesday, July 19, 2016 at 9:46:50 AM UTC-4, Roedy Green wrote:
> Oracle has renamed the files that form the Java API Java doc.
> 
> https://docs.oracle.com/Javase/8/docs/api/java/util/zip/ZipEntry.html
> is now 
> https://docs.oracle.com/javase/8/docs/api/java/util/zip/ZipEntry.html
> 
> Oracle's server is case-sensitive.

I already have a bookmark to docs.oracle.com/javase/8/docs/api/ so if this was a change it isn't recent.

[toc] | [prev] | [next] | [standalone]

#34980

On Tuesday, July 19, 2016 at 6:54:32 AM UTC-7, Eric Douglas wrote:
> On Tuesday, July 19, 2016 at 9:46:50 AM UTC-4, Roedy Green wrote:
> > Oracle has renamed the files that form the Java API Java doc.
> > 
> > https://docs.oracle.com/Javase/8/docs/api/java/util/zip/ZipEntry.html
> > is now 
> > https://docs.oracle.com/javase/8/docs/api/java/util/zip/ZipEntry.html
> > 
> > Oracle's server is case-sensitive.
> 
> I already have a bookmark to docs.oracle.com/javase/8/docs/api/ so if this was a change it isn't recent.

Likewise. I've had https://docs.oracle.com/javase/8/docs/api/ as the bookmark since the Java 8 API docs came out.

I observe that Oracle has not renamed the directories, let alone the files where the API docs reside.

-- 
Lew

[toc] | [prev] | [next] | [standalone]

#34982

On Tue, 19 Jul 2016 10:36:18 -0700 (PDT), Lew <lewbloch@gmail.com>
wrote:

>I observe that Oracle has not renamed the directories, let alone the files where the API docs reside.

Put it this way, I had all kinds of reference in my website with
Javase.  My link checker warns me of broken links after failing 5
days.  It went off this morning. Perhaps Oracle have just stopped
accepting Javase as an alternate.

[toc] | [prev] | [next] | [standalone]

#34983

On Tuesday, July 19, 2016 at 3:23:48 PM UTC-4, Roedy Green wrote:
> Put it this way, I had all kinds of reference in my website with
> Javase.  My link checker warns me of broken links after failing 5
> days.  It went off this morning. Perhaps Oracle have just stopped
> accepting Javase as an alternate.

It does seem odd.  I put these in Chrome:
http://docs.oracle.com/javase/8/docs/api/ is a valid web site.
http://docs.oracle.com/Javase/8/docs/api/ is not a valid web site.
https://www.google.com/ is a valid web site.
https://www.gOOglE.coM/ redirects to https://www.google.com/
https://yAhOo.CoM redirects to https://www.yahoo.com/

FTP paths would be case sensitive if the server is unix based.  I didn't think web sites were.

[toc] | [prev] | [next] | [standalone]

#34984

On Tue, 19 Jul 2016 12:49:02 -0700, Eric Douglas wrote:

> On Tuesday, July 19, 2016 at 3:23:48 PM UTC-4, Roedy Green wrote:
>> Put it this way, I had all kinds of reference in my website with
>> Javase.  My link checker warns me of broken links after failing 5 days.
>>  It went off this morning. Perhaps Oracle have just stopped accepting
>> Javase as an alternate.
> 
> It does seem odd.  I put these in Chrome:
> http://docs.oracle.com/javase/8/docs/api/ is a valid web site.
> http://docs.oracle.com/Javase/8/docs/api/ is not a valid web site.
>
If the underlying OS is case sensitive then, unless the webserver has 
been configured to be caseless, pathnames following the host name are 
also case sensitive. I'm not aware of any deviation from this.

> https://www.google.com/ is a valid web site.
> https://www.gOOglE.coM/ redirects to https://www.google.com/
> https://yAhOo.CoM redirects to https://www.yahoo.com/
>
But these examples are unsurprising because DNS is by definition case-
agnostic, so all three of these should 'just work' - and they do, 
regardless of whether you use ping or a webbrowser to test that.


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#34986

On Tuesday, July 19, 2016 at 4:32:08 PM UTC-4, Martin Gregorie wrote:
> If the underlying OS is case sensitive then, unless the webserver has 
> been configured to be caseless, pathnames following the host name are 
> also case sensitive. I'm not aware of any deviation from this.
> 
> > https://www.google.com/ is a valid web site.
> > https://www.gOOglE.coM/ redirects to https://www.google.com/
> > https://yAhOo.CoM redirects to https://www.yahoo.com/
> >
> But these examples are unsurprising because DNS is by definition case-
> agnostic, so all three of these should 'just work' - and they do, 
> regardless of whether you use ping or a webbrowser to test that.

I'm not sure what you mean, how that's any different, but it seems to me if I take any url and capitalize any of the letters I end up in the same place.  I tested this one for instance.

http://www.npr.org/2016/07/18/486464177/zero-days-documentary-exposes-a-looming-threat-of-the-digital-age

Oracle hosting their own site?  Possible someone at Oracle screwed up?

[toc] | [prev] | [next] | [standalone]

#34988

 Eric Douglas wrote:
> I'm not sure what you mean, how that's any different, but it seems to me if I take any url and capitalize any of the letters I end up in the same place.  I tested this one for instance.
> 
> http://www.npr.org/2016/07/18/486464177/zero-days-documentary-exposes-a-looming-threat-of-the-digital-age
> 
> Oracle hosting their own site?  Possible someone at Oracle screwed up?

What is the underlying OS for the NPR site?

Undoubtedly someone at Oracle did not screw up. If I'm wrong, I'll change my mind on the evidence.

-- 
Lew

[toc] | [prev] | [next] | [standalone]

#34990

On Tue, 19 Jul 2016 13:59:20 -0700, Eric Douglas wrote:

> On Tuesday, July 19, 2016 at 4:32:08 PM UTC-4, Martin Gregorie wrote:
>> If the underlying OS is case sensitive then, unless the webserver has
>> been configured to be caseless, pathnames following the host name are
>> also case sensitive. I'm not aware of any deviation from this.
>> 
>> > https://www.google.com/ is a valid web site. https://www.gOOglE.coM/
>> > redirects to https://www.google.com/
>> > https://yAhOo.CoM redirects to https://www.yahoo.com/
>> >
>> But these examples are unsurprising because DNS is by definition case-
>> agnostic, so all three of these should 'just work' - and they do,
>> regardless of whether you use ping or a webbrowser to test that.
> 
> I'm not sure what you mean,
>
If you don't know, look it up: either search for "caseless string 
comparison" of find the DNS RFC and read that. Since you apparently can't 
or won't search for it, in this case you want RFC 4343. All RFCs can be 
accessed at 

   http://www.ietf.org/rfc.html


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#35006

On Tuesday, July 19, 2016 at 6:03:53 PM UTC-4, Martin Gregorie wrote:
> If you don't know, look it up: either search for "caseless string 
> comparison" of find the DNS RFC and read that. Since you apparently can't 
> or won't search for it, in this case you want RFC 4343. All RFCs can be 
> accessed at 
> 
>    http://www.ietf.org/rfc.html

I have worked with unix before, I am aware of case sensitive systems.
I was not previously aware that a web server would allow case sensitive, as I'm sure I've never encountered one before.  Unless you want to obscure your publicly available data making it hard for anyone to get to without a link it doesn't make sense.

Either it's rare for anyone to host a web site on a case sensitive OS, or case sensitive OS normally use a flag that changes all urls to lower case?

I don't even know the point of a case sensitive OS as I've only seen it create problems.  If you have a unix server and a Windows client and you want to update program1 you have to make sure your file is not called PROGRAM1 and you can't rename PROGRAM1 to program1 on Windows without naming it something else first...

[toc] | [prev] | [next] | [standalone]

#35009

On 20/07/16 13:36, Eric Douglas wrote:
> On Tuesday, July 19, 2016 at 6:03:53 PM UTC-4, Martin Gregorie wrote:
>> If you don't know, look it up: either search for "caseless string
>> comparison" of find the DNS RFC and read that. Since you apparently can't
>> or won't search for it, in this case you want RFC 4343. All RFCs can be
>> accessed at
>>
>> http://www.ietf.org/rfc.html
>
> I have worked with unix before, I am aware of case sensitive systems.
> I was not previously aware that a web server would allow case sensitive, as I'm sure I've never encountered one before.

Have you never noticed that Javadoc is case sensitive?
How else would it correctly differentiate the case of classes/methods/members?

Try using a Javadoc classname URL which has incorrect case, e.g.
https://docs.oracle.com/javase/8/docs/api/java/util/Arrays.html
https://docs.oracle.com/javase/8/docs/api/java/util/arrays.html

> Unless you want to obscure your publicly available data making it hard for anyone to get to without a link it doesn't make sense.

Why would using case sensitivity make your data hard to find?
How often do you actually type in a URL (as opposed to just a website address)?

>
> Either it's rare for anyone to host a web site on a case sensitive OS, or case sensitive OS normally use a flag that changes all urls to lower case?

I suspect the majority of the WWW is hosted on case-sensitive systems.
The typical LAMP stack has case-sensitive filesystems, and Apache is case-sensitive by default.

Since you haven't noticed, it's probably safe to assume it's not a problem.

>
> I don't even know the point of a case sensitive OS as I've only seen it create problems.

A case sensitive fileystem is as important as a case sensitive language (both natural and computer). Case conveys information.
It doesn't cause any problems unless it's paired with a system which is not case-sensitive, or a user who doesn't understand it.

> If you have a unix server and a Windows client and you want to update program1 you have to make sure your file is not called PROGRAM1 and you can't rename PROGRAM1 to program1 on Windows without naming it something else first...
>

If you setup Samba correctly then you can rename the files. You can have multiple files with what to Windows would be the same name,
e.g. test.txt, Test.txt TEST.TXT are all possible simultaneously, visible to Windows and editable by Windows. You can also rename them.
Any problem is entirely the fault of Windows and its lack of proper support for case sensitive file names.
Linux does it properly. Samba attempts to do what it can but is constrained by Windows limitations.

I don't program on Windows so I don't know how javac on Windows manages Java classes which are not in files with the correct case filenames.
For example, suppose for some perverse reason you wanted the classes AbcDef and Abcdef. How would Java on Windows cope with the source and class files?

[toc] | [prev] | [next] | [standalone]

#35011

On Wednesday, July 20, 2016 at 12:11:08 PM UTC-4, Nigel Wade wrote:
> Have you never noticed that Javadoc is case sensitive?
duh
> How else would it correctly differentiate the case of classes/methods/members?
> 
> Try using a Javadoc classname URL which has incorrect case, e.g.
> https://docs.oracle.com/javase/8/docs/api/java/util/Arrays.html
> https://docs.oracle.com/javase/8/docs/api/java/util/arrays.html

Ever hear of a method called compareToIgnoreCase()?

> > Unless you want to obscure your publicly available data making it hard for anyone to get to without a link it doesn't make sense.
> 
> Why would using case sensitivity make your data hard to find?
> How often do you actually type in a URL (as opposed to just a website address)?

So you just assume no one will ever type in a url, and it's case will never change otherwise all existing links will be broken.  Of course the domain can't be case sensitive or phishers would be signing up for Yahoo.com, YAHOO.COM, etc.  Why would you need the rest of it to be case sensitive?  Do you have any examples of a url that gets 2 valid pages with different case?  Can you get 2 of the same email address with different case?  If they can find any domain case insensitive why should it be so hard to find a page on that domain with different case.

> >
> > Either it's rare for anyone to host a web site on a case sensitive OS, or case sensitive OS normally use a flag that changes all urls to lower case?
> 
> I suspect the majority of the WWW is hosted on case-sensitive systems.
> The typical LAMP stack has case-sensitive filesystems, and Apache is case-sensitive by default.

Majority?  Where do you get that statistic?  I tried changing case on NPR and it found the page.  I tried the same on ABC News and it found the page.  Case insensitive would seem the norm at least for news sites.  I would think most people would prefer case insensitive especially where there are shorter URLs their customers might be more likely to type in and get wrong.

> Since you haven't noticed, it's probably safe to assume it's not a problem.

I expect most of the time it isn't.  My company pays someone else to host our web site and they put it on Apache so we have the same problem with case sensitive.  This could be causing problems we don't even know about, if a potential customer typed in the url and couldn't find a page and gave up, but no one has mentioned any.  Just because it's not an apparent problem doesn't mean it's a good idea.

> >
> > I don't even know the point of a case sensitive OS as I've only seen it create problems.
> 
> A case sensitive fileystem is as important as a case sensitive language (both natural and computer). Case conveys information.
> It doesn't cause any problems unless it's paired with a system which is not case-sensitive, or a user who doesn't understand it.

Again I haven't seen any benefit of case sensitive files but I have seen it create problems.  What information are you trying to convey exactly with case in a filename, and what does that have to do with finding a web page without matching case?

> > If you have a unix server and a Windows client and you want to update program1 you have to make sure your file is not called PROGRAM1 and you can't rename PROGRAM1 to program1 on Windows without naming it something else first...
> >
> 
> If you setup Samba correctly then you can rename the files. You can have multiple files with what to Windows would be the same name,
> e.g. test.txt, Test.txt TEST.TXT are all possible simultaneously, visible to Windows and editable by Windows. You can also rename them.
> Any problem is entirely the fault of Windows and its lack of proper support for case sensitive file names.
> Linux does it properly. Samba attempts to do what it can but is constrained by Windows limitations.

Oh a Linux fanboy trying to put down Windows for never supporting a feature most everyone can live without.  I haven't seen Samba.  As I recall we just copied files from Windows to Unix with a basic ftp command.

> I don't program on Windows so I don't know how javac on Windows manages Java classes which are not in files with the correct case filenames.
> For example, suppose for some perverse reason you wanted the classes AbcDef and Abcdef. How would Java on Windows cope with the source and class files?

Without testing I would guess it gets an error because I don't type commands to compile Java and Eclipse doesn't accept it.  If I create a class file Test.java in Eclipse and put a declaration public class TEST { I get an error "The public type TEST must be defined in it's own file".  If I create a class file Test and select the New Java Class option and put in TEST I get an error "Type with same name but different case exists".  Where have you ever seen, or why would you ever want, 2 classes in the same package with the same name different case?

[toc] | [prev] | [next] | [standalone]

#35012

On 7/20/2016 1:39 PM, Eric Douglas wrote:
> [...]
> Again I haven't seen any benefit of case sensitive files [...]

     Because a case-sensitive file system spends no CPU cycles
converting letters to a "canonical" case, it is more efficient.

-- 
esosman@comcast-dot-net.invalid
"Don't be afraid of work. Make work afraid of you." -- TLM

[toc] | [prev] | [next] | [standalone]

#35013

On Wednesday, July 20, 2016 at 2:00:13 PM UTC-4, Eric Sosman wrote:
> > Again I haven't seen any benefit of case sensitive files [...]
> 
>      Because a case-sensitive file system spends no CPU cycles
> converting letters to a "canonical" case, it is more efficient.

That depends if you ever need to type your file names.  The inefficiency of mistyped names will likely exceed the efficiency gained from case insensitive matching, particularly on newer faster CPUs.

Windows Explorer also has a built in search option.  Does Linux have a file explorer with a search?  Does it have an option for case insensitive?

[toc] | [prev] | [next] | [standalone]

#35026

On Wed, 20 Jul 2016 11:08:53 -0700, Eric Douglas wrote:

> On Wednesday, July 20, 2016 at 2:00:13 PM UTC-4, Eric Sosman wrote:
>> > Again I haven't seen any benefit of case sensitive files [...]
>> 
>>      Because a case-sensitive file system spends no CPU cycles
>> converting letters to a "canonical" case, it is more efficient.
> 
> That depends if you ever need to type your file names.  The inefficiency
> of mistyped names will likely exceed the efficiency gained from case
> insensitive matching, particularly on newer faster CPUs.
>
Actually, it goes a lot further than that because it also applies to all 
filenames in system configuration files as well as those that are hard-
coded into programs, simply because nothing can guarantee that these 
filenames are in canonical form. Simply put, every time a file is opened, 
its name must be converted to canonical form, possibly rejected as 
invalid, and if valid, compared with the filing system to see if a file 
with this name exists. This is an essential prerequisite to checking 
access permissions to see of the program is allowed to open in in the 
requested mode.
 
> Windows Explorer also has a built in search option.  Does Linux have a
> file explorer with a search?  Does it have an option for case
> insensitive?
>
Several, including 'locate' (http://linux.die.net/man/1/locate) as well 
as the search tools in graphical file managers. Locate has a case-
insensitive option, the file manager I use does not.

Then again, the shells (equivalent to the Windows command-line 
interpreter, but rather more powerful - I use the bash shell, but there 
are several others) have a filename matching ability that can be used to 
provide case sensitivity if its needed, e.g. instead of using 'myfile
[0-9].txt' to select a file with a single digit before the extension, you 
could use '[Mm]y[Ff]ile[0-9].txt' if you're unsure whether the name 
prefix is myfile or MyFile. Here's a real-life example:

$ ls [Ee]tic[Kk]et*-[1-9].pdf
Eticket-DK7203-301791-4.pdf  Eticket-GL4002-301791-1.pdf

'ls' is the UNIX/Linux 'dir' command.
  

-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#35027

On 7/20/2016 2:08 PM, Eric Douglas wrote:
> On Wednesday, July 20, 2016 at 2:00:13 PM UTC-4, Eric Sosman wrote:
>>> Again I haven't seen any benefit of case sensitive files [...]
>>
>>      Because a case-sensitive file system spends no CPU cycles
>> converting letters to a "canonical" case, it is more efficient.
>
> That depends if you ever need to type your file names. [...]

     It never occurred to me that a smiley would be required on
so outrageous a claim.  Oh, well:

	Because a case-sensitive file system spends no CPU cycles
	converting letters to a "canonical" case, it is more
	efficient.  ;-)

(Sigh.)

-- 
esosman@comcast-dot-net.invalid
"Don't be afraid of work. Make work afraid of you." -- TLM

[toc] | [prev] | [next] | [standalone]

#35091

On 7/20/2016 1:59 PM, Eric Sosman wrote:
> On 7/20/2016 1:39 PM, Eric Douglas wrote:
>> [...]
>> Again I haven't seen any benefit of case sensitive files [...]
>
>     Because a case-sensitive file system spends no CPU cycles
> converting letters to a "canonical" case, it is more efficient.

And it may have mattered 40 years ago.

:-)

Arne

[toc] | [prev] | [next] | [standalone]

#35031

On 20/07/16 18:39, Eric Douglas wrote:
> On Wednesday, July 20, 2016 at 12:11:08 PM UTC-4, Nigel Wade wrote:
>> Have you never noticed that Javadoc is case sensitive?
> duh
>> How else would it correctly differentiate the case of classes/methods/members?
>>
>> Try using a Javadoc classname URL which has incorrect case, e.g.
>> https://docs.oracle.com/javase/8/docs/api/java/util/Arrays.html
>> https://docs.oracle.com/javase/8/docs/api/java/util/arrays.html
>
> Ever hear of a method called compareToIgnoreCase()?

Is that part of the W3C spec? I can't find any reference to it in any documentation regarding the WWW.
How would it be relevant to the fact that you didn't know that the WWW was case sensitive (the part of your message to which I was replying, and you have chosen to remove from your reply).

>
>>> Unless you want to obscure your publicly available data making it hard for anyone to get to without a link it doesn't make sense.
>>
>> Why would using case sensitivity make your data hard to find?
>> How often do you actually type in a URL (as opposed to just a website address)?
>
> So you just assume no one will ever type in a url, and it's case will never change otherwise all existing links will be broken.

I assume that anyone who manually types in a link has sufficient common sense to type in the URL correctly.
I further assume that people very, very rarely ever manually type in a URL. Links are followed directly from various document types, or they are copy/pasted. To enter a URL manually you would have to guess the endpoint of the link exists, and that's something that is almost never going to work. If you don't know what the URL is you can't type it in, if you do there is no excuse for doing it wrong.

The entire point of the WWW is that you follow links rather than type in URLs directly. Without links pages on websites are pretty much inaccessible.

If I were to change the case of any pages then I'd also change the case of any links which point to those pages.


> Of course the domain can't be case sensitive or phishers would be signing up for Yahoo.com, YAHOO.COM, etc.  Why would you need the rest of it to be case sensitive?  Do you have any examples of a url that gets 2 valid pages with different case?  Can you get 2 of the same email address with different case?  If they can find any domain case insensitive why should it be so hard to find a page on that domain with different case.
>
>>>
>>> Either it's rare for anyone to host a web site on a case sensitive OS, or case sensitive OS normally use a flag that changes all urls to lower case?
>>
>> I suspect the majority of the WWW is hosted on case-sensitive systems.
>> The typical LAMP stack has case-sensitive filesystems, and Apache is case-sensitive by default.
>
> Majority?  Where do you get that statistic?

Google.

>  I tried changing case on NPR and it found the page.  I tried the same on ABC News and it found the page.  Case insensitive would seem the norm at least for news sites.  I would think most people would prefer case insensitive especially where there are shorter URLs their customers might be more likely to type in and get wrong.


You've found two examples, and extrapolate that measurement to the entire WWW.
Not exactly a scientific basis for your thesis.

>
>> Since you haven't noticed, it's probably safe to assume it's not a problem.
>
> I expect most of the time it isn't.
> My company pays someone else to host our web site and they put it on Apache so we have the same problem with case sensitive.  This could be causing problems we don't even know about, if a potential customer typed in the url and couldn't find a page and gave up, but no one has mentioned any.  Just because it's not an apparent problem doesn't mean it's a good idea.

Why do you think any of your customers would ever type in a link manually?
What link do you think they might be guessing at, and typing in?
Most websites don't even have any kind of sensible links that you might ever have any chance to guess at, they are created automatically by CMS and CDN.
For example, this a current link from the BBC website:
http://www.bbc.co.uk/news/uk-36852365
(it's case-sensitve BTW, if you were curious).

and this is one from Amazon:
https://www.amazon.co.uk/HTML-CSS-Design-Build-Sites/dp/1118008189/ref=sr_1_1?s=books&ie=UTF8&qid=1469091672&sr=1-1

Why would you be guessing at links like that, and typing them in manually?

If you really want to know whether you are losing customers who manually type in random URLs for your website ask your provider to check the error logs for mistyped URLs where the error was simply down to incorrect case.

>
>>>
>>> I don't even know the point of a case sensitive OS as I've only seen it create problems.
>>
>> A case sensitive fileystem is as important as a case sensitive language (both natural and computer). Case conveys information.
>> It doesn't cause any problems unless it's paired with a system which is not case-sensitive, or a user who doesn't understand it.
>
> Again I haven't seen any benefit of case sensitive files but I have seen it create problems.  What information are you trying to convey exactly with case in a filename, and what does that have to do with finding a web page without matching case?

You were the one who brought up the issue of filename case sensitivity. Why are you now questioning its relevance to the topic you started?

>
>>> If you have a unix server and a Windows client and you want to update program1 you have to make sure your file is not called PROGRAM1 and you can't rename PROGRAM1 to program1 on Windows without naming it something else first...
>>>
>>
>> If you setup Samba correctly then you can rename the files. You can have multiple files with what to Windows would be the same name,
>> e.g. test.txt, Test.txt TEST.TXT are all possible simultaneously, visible to Windows and editable by Windows. You can also rename them.
>> Any problem is entirely the fault of Windows and its lack of proper support for case sensitive file names.
>> Linux does it properly. Samba attempts to do what it can but is constrained by Windows limitations.
>
> Oh a Linux fanboy

Resorting to personal insults is not the way to present a reasoned argument.

Rest of message ignored.

[toc] | [prev] | [next] | [standalone]

#35090

On 7/20/2016 1:39 PM, Eric Douglas wrote:
> So you just assume no one will ever type in a url, and it's case will
> never change otherwise all existing links will be broken.  Of course
> the domain can't be case sensitive or phishers would be signing up
> for Yahoo.com, YAHOO.COM, etc.  Why would you need the rest of it to
> be case sensitive?

Because path typical get mapped to file system that may be
case sensitive.

And because query typical get processed by some program that
may be doing a case sensitive compare.

>                     Do you have any examples of a url that gets 2
> valid pages with different case?

In real life no.

But you could easily create an example.

You could even create it in Java using Tomcat.

>                                  Can you get 2 of the same email
> address with different case?

The part to the right of the @ is a host name and is not case
sensitive.

The part to the left of the @ is system specific and can
be case sensitive.

I think it is less common that it actually is.

>>> Either it's rare for anyone to host a web site on a case
>>> sensitive OS, or case sensitive OS normally use a flag that
>>> changes all urls to lower case?
>>
>> I suspect the majority of the WWW is hosted on case-sensitive
>> systems. The typical LAMP stack has case-sensitive filesystems, and
>> Apache is case-sensitive by default.
>
> Majority?  Where do you get that statistic?

That the majority of web servers are using case sensitive file
systems is easy to back up.

https://en.wikipedia.org/wiki/Usage_share_of_operating_systems#Public_servers_on_the_Internet

How many of those are configured for actual case sensitive URL's is
a different question.

But it must be a really big chunk since it is the default.

Arne

[toc] | [prev] | [next] | [standalone]

#35088

On 7/20/2016 8:36 AM, Eric Douglas wrote:
> On Tuesday, July 19, 2016 at 6:03:53 PM UTC-4, Martin Gregorie
> wrote:
>> If you don't know, look it up: either search for "caseless string
>> comparison" of find the DNS RFC and read that. Since you apparently
>> can't or won't search for it, in this case you want RFC 4343. All
>> RFCs can be accessed at
>>
>> http://www.ietf.org/rfc.html
>
> I have worked with unix before, I am aware of case sensitive
> systems. I was not previously aware that a web server would allow
> case sensitive, as I'm sure I've never encountered one before.
> Unless you want to obscure your publicly available data making it
> hard for anyone to get to without a link it doesn't make sense.

It does not matter for links.

People typing in URL's manually are supposed to know that
while protocol and hostnames are not case sensitive then
the other parts of URL's may be.

> Either it's rare for anyone to host a web site on a case sensitive
> OS,

Unlikely. *nix is used for most web hosting.

>   or case sensitive OS normally use a flag that changes all urls to
> lower case?

It is possible. But I don't think it is common.

Arne

[toc] | [prev] | [next] | [standalone]

Page 1 of 2  [1] 2  Next page →

Back to top | Article view | comp.lang.java.programmer

csiph-web