Talk:The UNIX-HATERS Handbook
This is the talk page for discussing improvements to the The UNIX-HATERS Handbook article. This is not a forum for general discussion of the article's subject. |
Article policies
|
Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL |
![]() | This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||||||||
|
Change of opinion
[edit]Author of the book changed his opinion about UNIX as stated here: https://news.ycombinator.com/item?id=6234801 46.12.242.94 (talk) 00:51, 19 August 2013 (UTC)
- I think you're missing the point. The comment that it went from worst to best "without getting significantly better" is more an endictment of the terrible quality of most of today's operating systems. 97.86.156.248 (talk) 22:18, 22 December 2018 (UTC)
- Indeed. Many of the fundamental problems with UNIX are still present in the various versions and derivatives around today: The insanity of allowing ANY character in a filename except for "/" and NULL, the wildcard expansion done by the shell so that the program invoked cannot know exactly what arguments were supplied on the command line, the disastrous consequences which can result from a combination of those two problems, and so on. It always seems as though UNIX was a developmental system (which in many ways it was) which was allowed to "escape into the wild" before it could be properly tidied up into an acceptable, finished system (such as adding sanity checks to supplied filenames, for example).
- Some might argue that certain design decisions were a product of the time, which is partly true, but even then there were illogical decisions made in the original design which were carried forward through many versions. Why on earth choose the printable characters "@" and "#" for deletion? The system was developed using ASCII terminals which are quite capable of sending control characters (DEL/RUBOUT or BS/^H for character delete, CAN/^X to cancel line etc.). Sure these options can be changed, but why pick two printable characters as the defaults in the first place? And surely it would have been clear that using "`" (backquote) to enclose a command whose output is to be directed to the invoking program would inevitably lead to confusion with the regular "'" (apostrophe) character?
- UNIX was a mess from the start with its case sensitivity. Log in using all capitals and the system assumes a terminal which supports only upper case - A reasonable enough decision given the number of upper-case-only terminals around at the time. But to then translate everything to lower case? Combined with the ability to create filenames in which upper- and lower-case characters in the name are distinct, and the result is, to say the least, somewhat confusing. Just about every other system of the era stored filenames as upper case only, and translated lower case to upper case as needed when a terminal supporting both was in use.
- UNIX originally had absolutely no method of locking files or records within a file. It took years for that to be added, and then the methods diverged between System III/V and BSD. And don't forget that in some cases these are only ADVISORY locks, relying on a program to "behave itself" and not write to a file or portion of a file if the system says it's "locked." That effectively means that to employ reliable locking, one has to write separate routines for file access, and make sure the permissions are set so that only those routines can directly access the files, all other programs being required to call the appropriate process for file access. And this from a system which was supposed to be designed as a multi-user system from the outset? Let's not even get into the (un)reliability of locking across a network (which, to be fair, was likely not envisaged at the original development).
- The above points alone are enough to illustrate the poor design of the original system, without going into anything else. Yes, it has good points too, but it's quite incredible that UNIX and its derivatives still show many of these original design issues, as well as adding a whole load more. That the author came to consider UNIX et. al. to be the best available O/S without getting any better is a sad reflection on the way O/S design has gone since the book was written. 172.58.84.205 (talk) 23:25, 3 July 2020 (UTC)
Removal of 'External links'
[edit]It's not clear to me why this edit was made:
19:49, 27 November 2005 86.133.242.215 (→External links)
I removed one of the external links a day or so prior to that because it returned 404. The remaining external link is still valid, and, indeed, is a download page for the Handbook itself. As well, the stub tag was removed, and this article probably still qualifies as a stub, doesn't it? Am I missing something here? PaulHoadley 10:22, 27 November 2005 (UTC)
In the absence of an explanation, then, I've reverted that edit and restored the single functioning external link. I notice this has happened before, and I assume it's vandalism. PaulHoadley 02:56, 2 December 2005 (UTC)
Page move proposal
[edit]I believe (but could very well be mistaken) that the correct title of the book is 'The UNIX-HATERS Handbook', as opposed to 'The UNIX-Haters Handbook'. This seems to be inline with the 'UNIX-HATERS' mailing list that the majority of the messages within the book are taken from. If this is true (as it seems to be) Then we should move the page from 'The UNIX-Haters Handbook' to 'The UNIX-HATERS Handbook'. Thoughts? --Ryanfantastic 22:14, 25 June 2006 (UTC)
- Looks like we've reached a consensus! Thanks Marudubshinki for moving the page. --Ryanfantastic 05:35, 28 June 2006 (UTC)
- Aw shucks. All I did was read the linked PDF. --maru (talk) contribs 02:16, 29 June 2006 (UTC)
Original Research/Opinions
[edit]Added the tag because statements such as "both the intellectual content and the communicational tenor of the book are altogether so fraught with immaturity, envy, and utter nonsense" are clearly simply uncited opinions of the author unfit for an encyclopedia.Mahewa 12:10, 11 March 2007 (UTC)
- They've been removed. I'd just have nuked them if I were you, stuff like that isn't worth wasting time debating. Chris Cunningham 21:36, 14 March 2007 (UTC)
- Even when it's true, like this :p --bonzi (talk) 19:48, 10 April 2008 (UTC)
Could somebody please add a citing to the fact that the creators admit C, Unix were hoax?
[edit]It would be amazing if more people would know that the very creators of C and Unix ADMITTED that both projects were hoaxes and that Unix is one big fart. Not o mention everything else that followed it.
It would be nice if someone would also add this on the Unix Wikipedia page.
- Well, yes, they were alledged to have admittted this - but the date of the story is a clue :-) Snori (talk) 09:01, 11 February 2014 (UTC)
Thank you guys, this Handbook has opened my eyes to what a plague Unix really is. And I think I could safely say the same about everything Unix-like. — Preceding unsigned comment added by 92.29.173.20 (talk) 18:13, 24 October 2012 (UTC)
The book is now rather dated
[edit]"The book is now rather dated, most of the material being from around 1990, and many of the problems cited no longer exist..."
While this is true, it kind of misses the point being that many of the problems mentioned still exist in modern unix systems (after 20 years!). In fact, it is not beyond reason to expect the problems (overuse of c/c++, inconsistencies in the user interface, deficiencies of the shell) to exist in the next 20 years. I feel this deserves mentioning in the article. — Preceding unsigned comment added by 2003:50:AA11:377:1D0A:6003:CEEB:D237 (talk) 05:42, 6 September 2014 (UTC)
- Indeed, extensions and other developments might have solved a few issues, such as the lack of file locking, for example - something which should have been included in a multi-user OS right from the start, even if it couldn't do record locking because "everything is just a string of bytes." But there are still some fundamental design issues which exist to this day, e.g. the ability to create a file with a name containing just about any character except "/" - Even non-printable control characters! No sane OS should allow filenames such as A(!#jx%^=*)X - especially when certain characters are expanded by the shell before a utility or user program can even get to them. 97.86.156.248 (talk) 22:31, 22 December 2018 (UTC)
Old-school topic
[edit]Almost every free software and open source solutions are based on Unix-like systems, Linux-libre, GNU system, FreeBSD are really successful Unix-like projects of operating systems, you almost can't imagine using Git without an Unix-like system, due to this Unix-like success the Unix-heartbroken topics should be now old-school at all.202.40.137.201 (talk) 08:34, 27 December 2017 (UTC)
- Agreed. The idea that there is Windows and basically everything else is Unix-like is so ingrained in me (and I guess in most everyone millennial and later) that the "Many users had come from systems that they felt were far more sophisticated in computer science terms" part caught me by surprise. I still have no idea what better systems the article alludes to, which probably should be addressed. Especially if my suspicion is true that none of them are still active today. TheRustKitty (talk) 10:32, 4 June 2020 (UTC)
- Some operating systems that are mentioned in the book are VMS, Genera and ITS.
- I changed "in computer science terms" to "in features and usability", because the former is controversial and doesn't appear to be stated directly in the book. History has shown us that both Unix and C have survived and evolved for the last 50 years. Moreover, Thompson and Ritchie were awarded the Turing Award. Their accomplishments are both praised and unrivaled.
- Whereas the "Worse is better" article laments the absence of some features in Unix which do not seem to be necessary in retrospect. --Amakuha (talk) 18:35, 4 June 2020 (UTC)
Outdated material
[edit]Previously the article included more detailed outdatedness review, which I think is quite useful:
Many of the book's complaints about the Unix operating system are based on design decisions, such as the fact that the shell expands wildcards (which means that wildcards can only be used for filenames and not usernames, times or other things). For example, a large proportion of the complaints are about anomalies in the command line interface, and many of these complaints are still valid. But other complaints of the book have been addressed with time, such as the lack of a journaling file system and criticisms about the operating system's graphical user interface. The book predates the rise of Linux and thus concerns the several commercial versions of Unix then available (the inconsistencies between them being another major complaint in the book)—most of which have been replaced by inconsistencies between different versions of Linux and incompatibilities between various versions Linux, and between Linux and MacOS.
– I understand that it's unreferenced, but I think it's unfair that part of it was left in the article and part of it was deleted. --Amakuha (talk) 04:24, 1 May 2020 (UTC)