Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
On Thursday 03 June 2010 17:22:29 Oliver Albers wrote: > That's what my gut would tell me, yes. Nobody knows there even is a > difference. But nobody even cares. If that was the case, there wouldn't be this discuss about this bug. People who know the difference care, as it adds confusion when reading the FAQ. People who do not care will read the value provided in the FAQ and go with that, not looking twice. To me, it states K is 1 kilobyte - which is 1000 bytes. But then it goes on to say it is 1024 bytes. So as I know that is incorrect, what is the value of K? Is it a typo in the docs, and it is in fact 1000 bytes, or did they mean kibibyte? See what I'm getting at? > Willow:~ albers$ ls -lah foo > -rw-r--r-- 1 albers staff 1,0K 3 Jun 18:21 foo > Willow:~ albers$ ls -lab foo > -rw-r--r-- 1 albers staff 1024 3 Jun 18:21 foo > Willow:~ albers$ uname -a > Darwin Willow.local 10.3.0 Darwin Kernel Version 10.3.0: Fri Feb 26 > 11:58:09 PST 2010; root:xnu-1504.3.12~1/RELEASE_I386 i386 > > It is not just Windows. > > Olli 'K' is neither the prefix for kilobyte or kibibyte, so we don't know what they are meaning by that. That value could be rounded down (since it is only displaying to 1 precision).
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
> Are you trying to suggest that 90% of the desktop market share are unaware of > the difference? That's what my gut would tell me, yes. Nobody knows there even is a difference. But nobody even cares. > Also, is the desktop market (read, end users) going to care about the > difference or need to know? The FAQ in the documentation is not form them, it > is for developers. I can guarantee that 90% is not the figure we'd be looking > at if you gather statistics for the operating system of PHP developers. Willow:~ albers$ ls -lah foo -rw-r--r-- 1 albers staff 1,0K 3 Jun 18:21 foo Willow:~ albers$ ls -lab foo -rw-r--r-- 1 albers staff 1024 3 Jun 18:21 foo Willow:~ albers$ uname -a Darwin Willow.local 10.3.0 Darwin Kernel Version 10.3.0: Fri Feb 26 11:58:09 PST 2010; root:xnu-1504.3.12~1/RELEASE_I386 i386 It is not just Windows. Olli
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
On Thursday 03 June 2010 17:08:46 Oliver Albers wrote: > > As can be seen from the different file managers that I tried in response > > to Daniels comment; that isn't the common usage. The only file browser > > that was different was Windows Explorer. > > Which has a market share of about 90%: > http://marketshare.hitslink.com/operating-system-market-share.aspx?qprid=8 > > Olli Are you trying to suggest that 90% of the desktop market share are unaware of the difference? Also, is the desktop market (read, end users) going to care about the difference or need to know? The FAQ in the documentation is not form them, it is for developers. I can guarantee that 90% is not the figure we'd be looking at if you gather statistics for the operating system of PHP developers.
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
> As can be seen from the different file managers that I tried in response to > Daniels comment; that isn't the common usage. The only file browser that was > different was Windows Explorer. > Which has a market share of about 90%: http://marketshare.hitslink.com/operating-system-market-share.aspx?qprid=8 Olli
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
On Wednesday 02 June 2010 04:59:11 you wrote: > Given the number of supposedly competent developers who seem to miss > important bits of PHP manual pages as it is, I don't hold out as much > hope as you do on that front. > > Additionally, kilobyte is, in my experience, still the standard term > for "1024 bytes" in most technical environments. I just conducted a > quick (and admittedly extremely unscientific) poll of half a dozen > coworkers (a mix of programmers and admins), and when asked what a > kilobyte was, all answered 1024 bytes, not 1000 bytes. A more important question to ask them, is what value is 1 kibibyte? Remember, I'm not asking to change the value of K, simply the wording. They can still go on to believe that 1 kilobyte is 1024 bytes, that does not effect this. What matters is what 1024 bytes is, which is a kibibyte. If they don't know what a kibibyte is, they can read further in the FAQ or simply Google it. It doesn't add or take away any effort from those that do not know. Only adds confusion to those that /do/ know the difference. > Honestly, this seems like an advocacy issue to me, and I don't think > the PHP manual is the right place to be advocating the adoption of > specialist binary prefixes. I'd prefer to go with whatever common > usage is, and 2¹⁰ still seems like the common usage to me. As can be seen from the different file managers that I tried in response to Daniels comment; that isn't the common usage. The only file browser that was different was Windows Explorer. Everywhere that I have see, the correct wording /is/ the common usage. Regards
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
On 2 June 2010 01:21, Alex Cartwright wrote: > As a use case: Harry visits the FAQ to see what the value of 'K' means, it > shows that K means 1 Kibibyte. As Harry knows what a Kibibyte is, he knows > that 'K' should be 1024 bytes. Adam on the other hand, has no idea what a > Kibibyte is and so continues to read the FAQ which states that 1 Kibyte = 1024 > bytes (like the documentation currently does, albeit with wrong wording). I do know what a kibibyte is, thank you! :P (Sorry, couldn't resist. Moving on...) > My point being is that those who know the real value can continue to use the > real value; those that don't can either A) read a few words further or read up > on what a 'kibibyte' is, which takes less than 10 seconds. Something any > competent developer should be able to do. Given the number of supposedly competent developers who seem to miss important bits of PHP manual pages as it is, I don't hold out as much hope as you do on that front. Additionally, kilobyte is, in my experience, still the standard term for "1024 bytes" in most technical environments. I just conducted a quick (and admittedly extremely unscientific) poll of half a dozen coworkers (a mix of programmers and admins), and when asked what a kilobyte was, all answered 1024 bytes, not 1000 bytes. Honestly, this seems like an advocacy issue to me, and I don't think the PHP manual is the right place to be advocating the adoption of specialist binary prefixes. I'd prefer to go with whatever common usage is, and 2¹⁰ still seems like the common usage to me. Adam
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
> Windows, the most widely deployed desktop OS, agrees that 1 kilobyte = > 1024 bytes, so does Nautilus, the file manager in Gnome. I can't speak > for OS X, but I'm pretty sure it uses that definition as well. Nautilus 2.30.0 correctly reports a file of 1000 bytes as '1000 bytes' and a file of 1024 bytes as '1.0 KB' (which should be KiB but that's their issue). Dolphin 1.4 using KDE 4.4.3 also shows the same (correct) behavior. OSX Finder displays it as 4KB file due to the block size used, so that is of no help. The only one here that displays the same as the PHP documentation (as in, 1 kilobyte = 1024 bytes) is Windows Explorer. Which version of Nautilus are you using? As that is not what I am seeing > > Yes you can create what ever definitions you want, however what you > > should not do is change the value of those already existing definitions > > that are clearly outlined by standards. The International System of > > Units clearly, very clearly, states that kilo = 1000, not 1024. > > Indeed they do, but "byte" is not an SI base unit. SI is only > concerned with length, mass, time, electric current, thermodynamic > temperature, amount of substance and luminous intensity. Nor is "byte" > even included in the list of non-SI units for use in SI [1]. True however the value of 'kilo' still remains. > As Johannes said in the bug report, because it contradicts the most > widespread definition and thus contradicts what most users would > expect. Who are these most users? The majority of people I know understand the difference, and the ones that don't simply don't care or have a need to know the difference in them. As a use case: Harry visits the FAQ to see what the value of 'K' means, it shows that K means 1 Kibibyte. As Harry knows what a Kibibyte is, he knows that 'K' should be 1024 bytes. Adam on the other hand, has no idea what a Kibibyte is and so continues to read the FAQ which states that 1 Kibyte = 1024 bytes (like the documentation currently does, albeit with wrong wording). My point being is that those who know the real value can continue to use the real value; those that don't can either A) read a few words further or read up on what a 'kibibyte' is, which takes less than 10 seconds. Something any competent developer should be able to do. > Didn't see that you only replied to me. Please use "reply all" so the > mailing list gets a copy of the reply as well. Whoops, my mistake.
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
On Tue, Jun 1, 2010 at 18:37, Daniel Egeberg wrote: > On Tue, Jun 1, 2010 at 18:21, Alex Cartwright wrote: >> Not to sound funny, though where are you statistics for the 'commonplace >> definition' of kilobyte? Without that I don't think that is really a valid >> point to ignore such an obvious bug. > > This mailing list is not a scientific journal; I don't need to conduct > any research before I make claims. You're the one disputing the > definition in the manual, so you are free to collect statistics. > > Windows, the most widely deployed desktop OS, agrees that 1 kilobyte = > 1024 bytes, so does Nautilus, the file manager in Gnome. I can't speak > for OS X, but I'm pretty sure it uses that definition as well. > >> Yes you can create what ever definitions you want, however what you should >> not >> do is change the value of those already existing definitions that are clearly >> outlined by standards. The International System of Units clearly, very >> clearly, states that kilo = 1000, not 1024. > > Indeed they do, but "byte" is not an SI base unit. SI is only > concerned with length, mass, time, electric current, thermodynamic > temperature, amount of substance and luminous intensity. Nor is "byte" > even included in the list of non-SI units for use in SI [1]. > > SI does not have a monopoly on "kilo". I think you'll find the entire > English language (and indeed most other natural languages) "broken" if > you wish to assert that a word or a word prefix can only be used in > one single context. > >> I really don't see what the resistance is about of changing the word in the >> documentation from 'Kilobyte' to 'Kibibyte'. I'm not saying you need to >> change >> PHPs definition if 'K' at all, simply to change 'Kilo' 'Mega' and 'Giga' in >> the documentation to reflect what they really mean. > > As Johannes said in the bug report, because it contradicts the most > widespread definition and thus contradicts what most users would > expect. > > [1] http://www.bipm.org/en/si/si_brochure/chapter4/table6.html > > -- > Daniel Egeberg > Didn't see that you only replied to me. Please use "reply all" so the mailing list gets a copy of the reply as well. -- Daniel Egeberg
Re: [PHP-DOC] Bug #51834 set to bogus, your thoughts?
On Tue, Jun 1, 2010 at 17:31, Alex Cartwright wrote: > Morning, > > I reported bug #51834 a few weeks ago, and although I got a response which > agreed with me, it was set to bogus and I've yet to have any other response. > > How is it a bug reported by a user can be set to bogus even though the person > who did so agrees that it is a bug? It's only a small change of wording > required. > > http://bugs.php.net/bug.php?id=51834 It's set to bogus because it's not a bug, but rather a deliberate choice. "Kilobyte" meaning a 1024 multiplier is the most commonplace definition, so it's also the definition used in PHP's manual. You can create whatever definitions you want as long as you're consistent. According to Daniel's Standard Committee (which I founded two minutes ago and I'm the chairman of) defines one kilobyte to be the amount of bytes corresponding to Graham's number ran through the Ackermann function. I think we should start using that definition instead. tl;dr, there are conflicting definitions, you're upset that we chose the one you don't support and that we won't change it. -- Daniel Egeberg
[PHP-DOC] Bug #51834 set to bogus, your thoughts?
Morning, I reported bug #51834 a few weeks ago, and although I got a response which agreed with me, it was set to bogus and I've yet to have any other response. How is it a bug reported by a user can be set to bogus even though the person who did so agrees that it is a bug? It's only a small change of wording required. http://bugs.php.net/bug.php?id=51834 Alex