Linux Hamstrung by lack of standards? 70
Richard Finney wrote
in to send us a Yahoo/ZD article entitled
Linux hamstrung by lack of standards?
It mostly revolves around the Comdex stuff from monday, but
it has a few choice quotes, especially Larry Augustin predicting
that Linux will supplant NT in the server market this year,
comments about how Linux is growing in the corporate world,
and a little bit about the standards thing. Its odd that
they chose that title when the artice really doesn't
spend much time on the subject.
No, it's the Slashdot users that are idiots (Score:1)
Not a SINGLE criticism can ever be directed at Linux without people calling it FUD.
Is it too much to ask that there be a single agreed upon directory hierarchy and partition
layout in Linux?
Is it too much too ask that by default, Linux GUI applications all use common keyboard shortcuts, menus, and look-and-feel?
Is it too much to ask that Linux use a common configuration file format like XML rather than everyone inventing their own config file language?
When reporters talk about standards, they are not talking about IETF, ISO, W3C, or whatever standards. They are talking about having a standard experience on the desktop.
That is, you can go out and buy a "Learning Linux" book, get the concepts, and then use Slackware, Debian, Redhat, or whatever, without getting confused.
Standard APIs are one thing. But how about standard layout, standard keyboard, standard configuration and management, and standard style.
No one is taking away your god given right to have a freaky desktop and be a cool teenager, but atleast have some sensible DEFAULTS for the rest of the world.
Use whatever keyboard, toolkit, and layout you want (you can even choose non-standard install locations with InstallShield on windows), but the lack of a consistent set of guidelines and rules makes Linux extremely difficult to use by MOM.
For the realistic folks out that, let's face it. Slashdot users are foaming at the mouth raving religious zealots, just like Mac Advocates, and Team OS/2. Not a single bad comment can be made about Linux's shortcomings (how can you improve it if you don't realize it), without calling reporters morons, idiots, liars, borgs, FUDsters, or whatever.
Can someone please start Backslashdot, where mature nerds can get real nerd news without the OSS/GNU/Linux is sacrosanct bullshit?
Vague and Stupid (Score:1)
WIMP and sane interface design is what allows users to learn (assuming they wish to).
People interested in getting things done are just doing them and not whining. Hopkins FBI is sitting in my cubicle and CivCTP is sitting on my machine at home because it's the can-do's and not the cant-be-done's that drive Linux.
Linux could use a Direct3D 8.0 equivalent, done right and a 3D sound API. DirectX in it's current form really isn't anything to covet.
Why bother with 'what everyone will have'. If the libraries are free/gratic, the developer can just deliver them with the product (like that other OS does).
Serious DTP and pre-press is a more complicated issue and a matter of those problems not being dealt with as much on Unix (in general).
We certainly don't need to take the lead of Windows with high performance 3D graphics. (Unix created the widest & best abstracted standard for that)
Vague and Stupid (Score:1)
I wouldn't be surprised if they weren't even at the meeting and only grabbed a transcript to write up the article. I didn't see many reporter-looking people, taking lots of notes or pictures at the sessions. Aside from Linus' keynote of course, they were snapping pictures from it than leaving immediately, trying to make it around all the people sitting on the floor, standing on every open piece of floorspace they could get.
List o fud (Score:1)
I was there (Score:1)
Anyway, at this one, Larry Augustin was the first to speak, and that was how he started on where Linux will go in the next year, taking the server arena and a good bit of the embedded area as well. Sure, Ransom Love said we have to "actively promote" standards, but not in any way like Linux currently doesn't do that as the article sounds to me at least. We need to join together more than we are with all these various wars (distributions, desktops, etc). Have good, open standards, and if every distribution/desktop obeys them we have all the choice we have always had with compatibility across all Linuxes, not just Intel-based Linux.
And Bob Young was not there, but another guy from RedHat. Ah well, they don't pay complete attention to these details (or ask the men themselves, I would just to make sure I spell names correctly and everything, but that's just me). But the guy did say basically for all intensive purposes, you could call him Bob Young.
Yet another example of ZDnet FUD... (Score:1)
Yawn. Tired of mainstream Linux "news" (Score:1)
the biggest audiences have the least amount
content in their news. The only "homework" they
do is call someone on the phone and pester them
for quotes that can be taken out of context. Some
tech "journalists" will even call and try to put
words in your mouth and simply ask you to simply
confirm their brain-dead opinions -- I won't name
names but I've have some firsthand experience in
that area.
I like tech news stories where we can actually
learn something.
I've got your standards right here (Score:1)
Now let's look at Micro~1 Windows. Almost anything considered "standard" on Micro~1 Windows are things that are contrived by Micro~1 themselves. If you want standards compliance on Micro~1, you often have to buy things or install additional software.
Oh well. We all know the article was a waste of time anyway.
You don't get it... (Score:1)
Standards ARE NOT PROGRAMS. Here is a list of some common standards, and please don't flame me for what I say or don't say in this list:
ASCII
TCP/IP
X11
POSIX
the Win32 API
the i386 instruction set
CORBA
PostScript
Now, here is a list of some common things which are NOT standards:
vi or Emacs
BSD Sockets
XFree86
any Unix
Windows 9x/NT
the Pentium chip
KDE or Gnome
any PostScript-using device
What is the difference? Notice: everything in the first list is a protocol of some kind. Everything in the second list is an implementation of that protocol, or something which uses that protocol. Note that many different programs can adhere to a standard, each one being possible very different from the others in that category.
This is the thing: standards do not imply a lack of diversity. On the contrary, standards encourage diversity. Why? Because when different programs agree to a given standard, all of those programs can interoperate well, no matter how different they may otherwise be.
I'm going to give a rather controversial example to prove my point. Many people claim that Apple crunches down on consumer choice because it keeps dictatorial control of the standards used on the Mac platform. However, I beg to differ. Case in point: video cards. Without using a single driver, I can use cards from ATI, IXMicro, Mactell, Matrox, Micro Conversions, Village Tronic, and a whole host of others. Yet each of these cards comes with a "driver" to allow the card to perform even better. They are very different cards, and some do things that others do not, yet because they all adhere to the standard protocols defined by Apple, all of them will work well in a Mac, allowing for choice while following some basic rules.
This is what the article is complaining about. Mostly, it is wrong: Linux does have many standards in place. However, it is missing a few (mostly in the area of GUI's, but others do exist), and these are critical. I believe they can be nailed down without having to force any one program out of the market; they'll have to change slightly to agree to a standard but all of them will work, and a user will easily be able to swap one program out for another.
Here's the thing: (Score:1)
1) Configuration setups. I'm not talking about default configurations here, I'm talking about how a machine is configured. It differes somewhat across distros. This is a Bad Thing, the reason being that it means relearning configuration for each distro. This is not so bad when you only have a few distros out there, as Linux currently does. But what happens when every company suddenly comes out with its own configuration system?
2) GUI API's. X is not a GUI standard; it's a networking protocol on top of which a GUI can be built. This is often forgotten by Linux users (and users of Unix in general). X is also insufficient as a GUI standard. It defines the basic way to get an app running, but it doesn't define interactions between apps. Gnome and KDE are working on this, and they've made a very tiny bit of progress (XDND). But more standards than a drag-and-drop protocol are necessary. This can, by the way, be done without standardizing on one graphics toolkit (this last sentence started to hopefully avert Gnome vs KDE flamewars).
This is the thing: Linux has many standards hashed out. But what it doesn't have is very important.
And another thing: ZDNet seems to subscribe to the M$ belief that an application can be a standard. This simply isn't so; protocols and designs, not implementations, are the true standards.
List o fud (Score:1)
Looks like I will quite busy as the fud seems to fly faster these days. Searching microsoft's site turns up some gems.
List o fud (Score:1)
Headline is /. bait! (Score:1)
Guys, wake up and smell the roses!
This headline is pure slashdot bait. They want to get the slashdot effect, so they get more hits, so they get more $$$$$ from the advertisers!
I'm sure a load of news sites have cottoned onto the fact that a good, juicy Linux headline gets them at least 10,000 hits extra. Rob should set up a scoring system, whereby each post about a "mainstream" news site reporting on Linux (or any other OSS topic for that matter) is given a relevancy score. That'll save us all the time it takes to trundle off to ZDNet or News.com to read yet more uneducated rubbish.
--
Barry de la Rosa,
Reporter, PC Week (UK)
Work: barry_delarosa[at]vnu.co.uk,
tel. +44 (0)171 316 9364
Clickthroughs? -- dream on (Score:1)
Sorry mate, you're wrong here. Selling ad space based on clickthroughs is a suckers game. CPM (as anyone in the print business will tell you) is cost per thousand
Would you sell space for $15 per thousand clickthroughs? You'll be waiting a long time to get that first $15...
Plus, ZD probably charge more than $15 cpm - it's a prime site. I think you'll find it's worth their while.
--
Barry de la Rosa,
Reporter, PC Week (UK)
Work: barry_delarosa[at]vnu.co.uk,
tel. +44 (0)171 316 9364
Richard Finney Hamstrung by Lack of Intelligence (Score:1)
hehehe
Can anybody be so blind (Score:1)
I just hope that enough people who actually develop software aren't as ignorant as you.
There is a point to it, however. (Score:1)
However, Linux is now beginning to struggle in the desktop environment. It is my opinion that open source development will only now be put to the test. Let's face it, as of now I cannot copy and paste in X windows. Sure there are a slew of programs the work together, kde, gnome. But what about the thousands of other programs. Should they just remain at odds with the rest of the desktop?
Nobody in the Linux community likes standards. Why? Because a standard is by definition NOT open source. Anybody and everybody cannot change a standard when they think that it needs tweaking or else it is not a standard. This idea frightens GNU maniacs. It is my challenge to the Linux community to suck in our pride and declare certain protocols as the best in theory that can be made. Then finalize them.
Perhaps we can then finally get a standard way to play sounds, video, games on the console, x windows GUI continuity.
This isn't to say that we have to abide by one snippet of code. But rather a protocol. To play a sound you open an IPC to the registered sound daemon. If that were specified, it doesn't mean you can't switch that sound daemon for a faster one. As long as they work the same at the surface level.
Damn straight (Score:1)
There is a point to it, however. (Score:1)
Neither could I , until I read the manual that is
What server market are they talking about?? PC? (Score:1)
Seems most of these experts forget about all the mini computers out there that perform server duties, and do them well.
Microsoft is trying to position their NT software into the Enterprise category, it won't work there very well. Its hamstringed by the fact its being written as a single user product... iow - a product designed to run on a seperate pc....
Any server product whose design includes being run on a singular machine won't make it to the enterprise level.
No, it's the Slashdot users that are idiots (Score:1)
Where do you find anything in the Windows documentation that equals the HOWTOs? Windows documentation is for newbies, "power users" are supposed to buy a "resource kit".
No, beginning of story. Do you really see the Windows interface between applications as consistent?
The list goes on.
Only UNIX can give you standards (Score:1)
Furthermore, while many of the base APIs in Unix are old and stable, the modern, more concerted efforts to create an integrated environment are not. Ten years from now, it may be technically possible to run your legacy Gnome 1.0 application on your Pentium64 SuperPro, but it could involve finding and trying to compile all 30 legacy Gnome libraries. The cost/benefit of porting starts to fall of really quickly.
There's a few more dependancies involved other than just "Unix is Unix". When GNU refused to port gcc to Apple A/UX (System V-based with MacOS GUI), it crippled the platform. All the wonderful Unix source code available is hardly any good when you can't compile it. Future changes to gcc could render quite a bit of existing Linux code obsolete.
Sure the Windows API is a moving target. On the other hand, most Win16 and non-game DOS programs run just fine on WinNT4 and probably also on Win2000. This binary compatiblity has given Windows quite an advantage over Unix's source compatiblity on the desktop (at the cost of the WinTel duopoly).
Linux standards are an opportunity to change all of this. Let's hope RedHat, Debian and so on can play nice and avoid the small picture thinking and proprietary maneuvers typical of other Unix vendors.
--
Vague and Stupid (Score:1)
One thing that always gets to me when people talk about linux, or, really, anything, is the artificial generation of limitations. If there's one thing both the GNU tools and the linux kernel are based on, it's standards.
Maybe what they meant to say is, "My company has an agenda. Part of that agenda is the adoption of certain methodologies and naming conventions. Unless I make you all afraid that not adopting these conventions will hurt linux, my company is less likely to post large profits, and I may not be able to afford my new boat. Please post lots of comments about how linux needs to conform to my standards, in order that my 5 year old child might be more assured the joy of killing numerous large fish this summer. Thanks."
Here's a clue, pal, and folks. Linux does not 'need' anything but to keep doing what it does. Linux does not 'need' adoption by anyone. Whether the media trumpets its every google and spittle or not, it will continue to rise, and continue to rule.
--
MS has 22 billion dollars in cash, and none of it is mine.
Vague and Stupid (Score:1)
One thing that always gets to me when people talk about linux, or, really, anything, is the artificial generation of limitations. If there's one thing both the GNU tools and the linux kernel are based on, it's standards.
Maybe what they meant to say is, "My company has an agenda. Part of that agenda is the adoption of certain methodologies and naming conventions. Unless I make you all afraid that not adopting these conventions will hurt linux, my company is less likely to post large profits, and I may not be able to afford my new boat. Please post lots of comments about how linux needs to conform to my standards, in order that my 5 year old child might be more assured the joy of killing numerous large fish this summer. Thanks."
Here's a clue, pal, and folks. Linux does not 'need' anything but to keep doing what it does. Linux does not 'need' adoption by anyone. Whether the media trumpets its every google and spittle or not, it will continue to rise, and continue to rule.
--
MS has 22 billion dollars in cash, and none of it is mine.
So.....? (Score:1)
Headline is /. bait!? -- dream on (Score:1)
2. Clickthroughs are the wobbly foundation of web site cash flow. The industry average is around $15/M and going down. Clickthroughs per pageview on many sites are below 1%. Do the math. 10,000 page views *might* be worth a few dollars.
3. You work in this business and you believe ZD would try to build traffic this way? Even if they could, never attribute to conspiracy what you can attribute to coincidence.
-------
And Windows 2000 removes legacy support... (Score:1)
Yet another example of ZDnet FUD... (Score:1)
Next year, who knows...
It was really interesting that I could get my new homebuilt machine to work w/ Linux before NT and Win95. Something about having to download drivers anyway since the one's shipped w/ Creative's TNT are broken. Man, if vendors keep this up, Linux will have world domination in six months.
Take care.
Reporters "Hamstrung" by Linux (Score:1)
Most reporters don't understand Linux. Their only base of knowledge is the Microsoft world where Microsoft feeds them the sound bites they need to write their articles. They don't have this with Linux. Everyone is openly discussing the problems and potential problems of Linux without a single source to put a positive spin on things.
Yet another example of ZDnet FUD... (Score:1)
It really is too bad we can't hold them to British libel standards.
"... which is why we're going to take over the world." -- Linus, 4/12/1999
Mutual exclusion is the problem (Score:1)
While I agree that unix (especially free unix) thrives on standards, I think you're not making a critical distinction here. What we have are two different notions of 'standard'
To the corporate types a 'standard' is a dominant product in the industry that they can feel safe with because they know it won't be going away any time soon. This is why some people consider microsoft to be a 'standard'
To the unix geeks a standard is some kind of protocol or api that allows for programs to interoperate, i.e. posix, X, etc. This is largely independent of implementation, as is evidenced by the plethora of unix variants that will all run the same C programs pretty much right out of the package. This happens in the free world as well and it is a healthy thing.
I think that the gnome/KDE split (while excessively nasty at times) is a virtue. Our community thrives on parallel development and adherence to standard interfaces. While a KDE program might use qt and a Gnome program will use gtk, the differences pretty much stop there, as it is looking like kde and gnome will talk to eachother. These are not mutually exclusive software packages, so neither of them needs to become a 'standard' in itself. As long as they can interoperate there are no problems.
What about the duplication of effort you might ask? I say that is healthy as well. It seems that when code is useful to all it is shared, and when people have different ideas as what to do they do things differently. If the idea works out well, others will use it. Imposing any more structure on it than that is unnatural and can't be done.
Even when different software packages are mutually exclusive, for example the linux and freebsd kernels, it is improper to say there are no standards because while kernel internals may differ there are well established interfaces that portable code can utilize to avoid platform dependence.
And so what if joe user doesen't care about diversity. Free software is not driven by joe user. Let the masses run redhat. Redhat is no reason debian can't continue to exist. Free software is a different paradigm where the traditional corporate rules just do not apply. Free software is interoperable because it is in its best intrest to be so. Yes some distributions will die. But the good ones will not, because if they are good then people will realize this and pick up the project if the origional author leaves off. This has happened to so many projects it's not even funny. Gimp/Gtk is a good example. So is the debian distribution. It's not like when a company dies and takes its products with it.
Only UNIX can give you standards (Score:1)
By contrast, UNIX code from 1972 would run unchanged in 1998. 10 year old code just keeps on running. 25 year old code too.
I forget how old X is, but it's pretty old and still backwards compatible from way way back. An old OpenWindows program will still run with no problems.
Having said that, yes standards are important. Personally I think the Gnome vs KDE divide is a massive problem. Joe User doesn't care about diversity. He wants to be able to take a shrink wrapped package (or rpmed package) and have it fit in with his desktop. He's not going to be happy if he's using Gnome and the program he wants to use only works half right because it was designed for KDE (or vice versa).
I also don't see a problem with one distribution dominating provided they do a good job. I don't see the need for so many. I mean it's just putting a whole lot of free code together and making sure it all works together. It has to be done, but it's not really productive for 10 different groups to all be doing it. They'd be better off building something new and useful.
KDE or Gnome will die. Some of the distributions will die. It's a shame to see the code go to waste, but it's better than the alternative - bad interoperability.
And Windows 2000 removes legacy support... (Score:1)
How many customers will spend how long waiting for vendors to port stuff to Windows 2000 now...
headlines lie (Score:1)
Or maybe that's as far as the editor got through the article before he/she decided on a headline.
God, I am getting sick of this!
Headline? HeadLIE!!! (Score:1)
Really, that article did have a pretty innacurate title...
No, it's the Slashdot users that are idiots (Score:1)
No, it's the Slashdot users that are idiots (Score:1)
Oh Contraire (Score:1)
The Problem with Standards is... (Score:1)
>this is what standards were originally about.
>If we can get consesus about things, great. What
>we don't need is another "standard".
Consensus is a good way of putting it, but for consensus to work properly, it must be documented somewhere so it can easily be referred to.
A imposed standard is contrary to the "best strategy will win" basis in linux, and it's free forming nature, but a consensus is a result of free software, where multiple idea can, and will, converge on the best solution naturally.
To me, standards like the FHS, etc, are like the "order" part of "law and order".
For example, there is nothing fundmently wrong with driving on the right or left of the road.
However, because of the number of users, one side must be chosen, not because it's "correct", but because one must be chosen to avoid chaos.
On the flip side, there are other places where standardising on one idea is a bad thing, and could lead to stagnation. GNOME and KDE have lots of "rivaliary" between them, and I think the speed of development results partially from that.
Of course, the hardest part, is deciding which parts need to be constant, and which parts need to have competition to spur them on.
The most interesting thing to come out of Comdex.. (Score:1)
Bill Gates is a sneaky bastard, after all and he knows that numbers (preferably numbers preceded by dollar signs and followed by percent signs or words like million and billion) attract the attention of the PHB crowd (who, as we all know, are the idiots who make the buying decisions which productive individuals then have to live with).
ZD fud doesn't matter.
Gates not "getting it" doesn't matter.
All that matters is that Linux remains what it is and has been: A better solution. So long as we don't lose sight of that, it won't matter how much fud gets spread around, because it won't do any good. The geeks are going to win.
--=+=--
Jesu Christi Dominatrix!!
DoktorMel
So how do we change the news? (Score:1)
But I want to know this: How do we change the news?
Do we e-mail articulate replies to the editors (instead of crashing their servers)? Do we send in our own well-written, thoroughly proofed and edited articles? (Not technical FAQs, but articles for mainstream readers and Micro$oft clones.) What is the best way to help technical journalists to produce accurate articles about Linux? One of the reasons why Micro$oft gets wonderful reviews is because it helps writers, editors and advertising directors in SO MANY WAYS$$$$.
What can we do? Appeal to honest, ethical reporters. (Yes, there are some.) And expose those "journalists" that are not.
Zd's lack of standards (Score:1)
I should know I work for them !
The Problem with Standards is... (Score:2)
Today, standards are a way similar to patents that companies are using to compete. I am not all that hot on Standards.
But what I think is important is Consensus...this is what standards were originally about. If we can get consesus about things, great. What we don't need is another "standard".
ttyl
Farrell
Standards (Score:2)
Well, other than System V, POSIX, Unix98, some BSD, net standards drawn up by the IETF, stuff like that.
But, you see, those don't count, at least not to journalists, as they're not =Microsoft= standards. The fact that the REST of the planet uses them is quite irrelevent. (For example, Multicast routing, IPv6, RSVP, QoS networking, etc, don't officially exist, as no Microsoft product is capable of supporting them. How can they exist? If MS doesn't use them, there can't be any standards, so clearly they don't exist.)
Vague and Stupid (Score:2)
This person, anonymous and cowardly tho they may be, has hit on exactly the point of my comment. To hell with the corporations and their vaporous standards. Look around, see if anyone else is doing what you want to do, and, if there IS a standard, code to it. If not, hack, hack, hack, and fix it later. And, what the hell is this about lack of gui standards? Do any of the people who say things like this actually program? Do they know any 'developers?' Enough of the buzzwords! Shut up and code, or shut up and use, or shut up and write a damned RFC for your magical missing standards.
Standards (Score:2)
The question is can we make enough money to justify the development and support costs.
Also ask whether you can make enough off of all Unix versions. Most of the APIs a typical app would use in Linux are not Linux APIs but standard APIs supported on almost all Unices. If it compiles on Linux and you didn't make assumptions about things like byte order and word size and stuff like that, there's a good chance that, with minimal work, the same code will compile on Solaris, AIX, HP-UX, Digital Unix, BSD etc. etc.. Tools like GNU autoconfig can help by doing the scut-work of figuring out what is supported on the platform you're compiling on this time so that, instead of writing code to platform X, you write based on "of capabilities X, Y and Z that would do the job, which are available on this platform?" ( eg. does this OS support select(), poll() or both? ). The GNU tools are prime examples of code that is designed to compile on multiple platforms, as some of those programs support literally an insane number of CPU/OS/version combinations.
The next question is do we try and support all the distributions, or just one (Red Hat).
Again, often you won't care about distribution. For example, for the C library calls, the question isn't "Is this RedHat or Debian?", but "Do I use libc5 or libc6/glibc2?" and "Do I need glibc 2.1.x or will 2.0.x work?". If, for example, glibc 2.0.7 works, then the same program should run fine on any distribution that uses glibc2 including RedHat and I believe the latest Debian.
File locations may get trickier, but the FSSTND and FHS are good guides. Use config files or environment variables to specify file locations ( with compiled-in paths based on FHS/FSSTND as fallback ), or depend on Unix facilities ( eg. use PATH via exec[l|v]p() or file-existence search to find executables instead of hard-coding their location ) and you can make adapting to a distribution's layout a trivial matter of making sure PATH is set correctly and a few lines in a config file are edited appropriately.