Who cares about Perl? It's gonna dyi anyway soon along with Python. The result of their merge, new language Parrot will be very unstable for awhile, uncompatible to both Perl and Python. So any investment of time, efforts and books to Perl (and Python) will be lost.
I think it's because of lack of standard of both Perl and Python.
If Perl is dying - where to go? Ruby? Erlnag? Lisp? I stick with Lisp for while. At least it's very clean, very clear and has a lot of libraries - that's exactly what I need from a language for web management.
"suitable ruleset configuration" - this is the key. The system (change and configuration management) should be rule-based. The reason why all previous open source (and even commercial) CCM systems failed - their architects ignored results achieved in AI research and, therefore, they made very unflexible systems.
Based on my software development expierience (17 years) I can tell, that there is no clear borders between models of software development process organization. Every team is uniq. Therefore, there is no universal models. But there are rules:
Rules to control access (based on ID, groups, roles, files, directories, branches, tags, schedules etc).
Rules to replicate distributed repositories. Rules to roll-out new versions.
Rules to resolve conflicts
Even rules to create "smart" diffs
and, of course, rules to "report" (if CVSWeb would be made with XSLT).
Prolog may help, but it's not a general programming language. Lisp is good, but it is not popular. Yet. Although, there is at least one project on Lisp related to CVS improvement: Meta-CVS. Python is much more popular and can be used effectively for AI programming, but there is no official standard for Python and that's why it is gonna dyi soon. Together with Perl. Erlang? Dylan? Too proprietary and there is not enough of libraries. The other languages are not good for functional programming. That's why I would support the effort of Meta-CVS, which is based on Lisp. Just I would go further and implement complete rule-based distributed change and configuration management layer over CVS.
CVS already have triggers. Multi-site repository requires better access control and conflict resolution, probably both should be rule-based. But it sounds like workflow management, which is huge development - there is no good workflow management done open source, is it?
Quicktime is useless until it supports Linux, *including* linux86, linuxppc, linuxalpha and linuxsparc. I guess, better to have a compilable-on-gcc source code yet.
I see the history. All article is about the past. Where is future?
Correct me if I am wrong, but I thought about the future of 2.0 in terms of design and features of the version greater than 2.5 and, maybe, of 3.0. What is planned in future releases?
As for 2.0 itself - who cares about the dead meat. We must use 2.4 or 2.5. Period
It only took me 5 seconds to find my other computer in Windows, but I haven't found it yet in Linux. It does need to get a little bit better.
It's not just linux - on any *n*x it's the same problem. The reason is the lack of something simular to NT domain and transparent network security.
More technically - OpenLDAP is supposed to solve it, but all vendors ignore it. All (almost) *n*x vendors still use obsolete (although often shadowed) password files in/etc instead of authentifying through OpenLDAp, which is is usually included to distro, but as something alike "contrib", for info, not for use. Without LDAP - no network-transparent authentication, network-transparent user-group security and thus - no secure way to share forlders.
Another general comment. *n*x came as a server - all users login from terminals to the same server where they share their folders under the same root "/". What is that "root" in network? Don't mention me already-depricated NFS - it's not for modern tasks. It's not reliable. And it's not secure.
Linux today is just coming to desktops (remember days of M$-Win 3.1?). Groupware environment is down the road. Be patient.
I remember C-M-+ in X11 documentation from the time used 0.95 version of Linux kernel - that means very long time ago. And that also means the guy is lazy to read docs.
So lazy that don't use google to find anything about "linux X11 resolution", which point to lots of howtos and docs, including classic ones on www.linux.org and Linux Documentation Project.
Lazy novices should be among top ten reasons why good projects fail.
In Python, using functional programming, you can work with XML on a very high level, like RDF and ontology. Check it here.
In Perl all you can do is moslty a primitive old-style regexp-oriented programming. Sort of grep or awk against XML files.
The reason is simple. Python as a programming language that has been designed by programmers. Perl is just a bad luck of its after-awk evolution done by programming system administrators.
The most densive way to encrypt the information is DNA:
It works very stable for billions of years;
It doesn't allow "bad" or useless information to exist too long;
It's already known to modern human science;
Create some plants (I'm a vegetarian), better trees, encode all libraries to their DNA in order for information to be reproduced on tree leaves as pictures of texts with diagrams, better in English alphabytes, but Chineese or Cyrilic will work as well. Distribute such plants between public parks with a condition to grow such trees.
Imagine - you go for BBQ and read some archived slashdot articles right on the tree:) Well, Slashdot might be wiped out by the evolution. But Einstein's articles will certainly stay forever! I am sure. Otherwise, why would the evolution allow Albert Einstein from the first place?
I can afford G4 hardware, but I cannot afford cheesy snobby UI of Mac OS X. If it's BSD-like system - give me my X11! Give me the choice of destop and window managers! No? Keep it away from me. I install Linux/PPC (Yellow Dog Linux this month, later it might be Debian or Suse or one of other linux distros) and enjoy.
I enjoy the choice, I enjoy the development environment, I enjoy the ability to test my software in different GUI styles. I enjoy hte performance of my PostgreSQL, Zope, Tomcat, gcc compiler and various scriptings. I don't want buggy OS. I don't want stability and compatibility problems. I want the stability, compatibility and my freedom.
What about Mac OS X? It's the same crap as M$ winz. No future.
Three-Dimensional Workspace Manager, a Chalmers Medialab project, released as Free Software under the LGPL license and focusing on the development of a general-purpose platform for 3D user interfaces (3D GUIs). 3Dwm is NOT an X11 window manager and cannot work as such. It is a full 3D user environment, i.e. the 3D equivalent of X11 (far from completed, however).
Instead of confining your applications to the conventional 2D desktop, 3Dwm displays them in three dimensions and also provides the necessary for creating new applications with full 3D-UIs.
Designed especially for use in the Chalmers Medialab 3D-CUBE, 3Dwm will also run on most consumer-level desktop systems as well.
If AOL can't even make ICQ and AIM interchangeable.
With Jabber on my Linux box both AIM and ICQ are interchangeable. If AOL will publicly ask for help I will do that. And of course a lot of other Jabber users can help:)
Both Perl and PHP are not really OOP lanuages. Both GUI and web tasks are UI at the end and therefore require OOP. Use Perl when you need write-only regexps readable only by maniacs. Use PHP for small pages, where you have to use programmers with VERY limited IQ. For UI - use either Java or Python or CLOS/Lisp. Or even C++, which is extremely unportable and thus I don't consider it as a choice.
If you hate Java as unscalable proprietary lang then use either Python or CLOS/Lisp.
For those who argue about security of object attribute access - that has nothing to do with OOP and with security either. It's just a marketing blob and nothing useful. Hence, Python and CLOS/Lisp are still the choice.
Finally, CLOS/Lisp has not enough portable GUI libraries. Scheme is not OOP lang either. Thus we've left alone with Python. Portable open-source scalable stable-as-old and modern-as improved good OOP (and some FP!) lang Pyhton.
I successfully compiled Icewm, AfterStep and WinMaker. And then discovered the reason why they are not inlcuded into the distro from the first place - they are way too buggy. They manage windows well (well, otherwise they are not window managers), but all those extensions - workspaces, lounchpads - are just broken. At least on my win2k box.
As for UNI*X features not yet available in Cygwin - you mean Gnome ported to win32, do you? At least that what failed the compilation Sawfish and others for me.
Have anyone heard about porting Gnome to win32? How about KDE? I can use Gnome or KDE through X11 net, but how about the case of temporary standalone box? Such win32 port would be a very nice feature then.
Open Source systems have a lot of alternatives. Big ones, like KDE vs GNOME, and small ones, like various PHP scripts. All this Open Source soup is evolving. And same as the mother-nature, it does it slow without sex (without exchange of source code pieces), or rapidly with sex (with exchange of source code pieces). Got a point? We have to steal software source code from each other in order to accelerate our software evolution!
All those themes change almost nothing - it's not a real customization I want as a user. And it's not a level of customization OEM would like either.
Finally, it's not scalable down - it requires a lot of hardware resources and cannot work on low-end computer. It cannot be scaled up either - how about network application server? With both *n*x/X11 and M$/VNC it's easy.
Conclusion: Mac OS X is just a default OS installed on last Mac computers to make them more attractive within the sales process. But we all know that atractive-for-sales features are often different from useful-for-real-work features. Mac OS X is just a sales-toy. The more Linux is improved is the more often Linux will be installed after-sales.
Mac OS is the reason of low sales of Mac computers. Mac OS X is a pitty attempt to fix it. But only with modern Linux (pre-installed!)the situation with Mac sales can be improved.
When Steve Jobs will uderstand it, the production cost in Apple Corp wil be dropped and the sales will jump up.
I heard that Bill Gates payed ~$100M to Apple specially to keep them away from Linux. Isn't it true?
when I say it's quarter to ten, I wouldn't want to see it represented as the fiddly non-integer 9.75.
If an hour would be equal to 100 minutes, then a quater to ten will mean 9:25, because in metric system a quater means 25%. Did you had a chance to use 25 cents (quater) coin?
As for third... do we even really need any thirds?
Linuxconf's major design problem is its own config database, which is used to generate real config files. On practice it drives to inconsistency to your own manual changes or they will be eliminated. Besides, you don't have good control over the generation process and linuxconf sometimes can regenerate some of config files even you didn't mean it.
Webmin is much better tool at least in its design. It reads real config files, so your manual changes will never be lost. It's still premature, though - some of modules, like usermin, are broken.
The major design problem of all such systems is a lack of versioning, (ACID) transactions and unified (XSLT?) reporting mechanism.
There is no pressure on that side from alternative vendors (Apple, microsoft), so I don't expext any big improvement here soon.
The best commercial admin env in Unix I saw was Smith in IBM AIX. It's object-orinted (sort of), good designed, and most important - it helps to learn CLI administration by giving the two-way access to CLI commands it runs. I wish IBM would contribute it to Linux at some day.
I agree. Apple doesn't have unique features. It has unique design. And that is not for Japan, as far as I see from the original article.
Wait. Apple has one unique feature - it's its hardware PPC platform. Wich works even better under Linux/PPC, which is more about Linux rather than about Apple.
Slashdot Mirror
It is compatible with linux... But the question is - can I buy it *WITHOUT* M$ crap and without paying for the M$ license?
I think it's because of lack of standard of both Perl and Python.
If Perl is dying - where to go? Ruby? Erlnag? Lisp? I stick with Lisp for while. At least it's very clean, very clear and has a lot of libraries - that's exactly what I need from a language for web management.
Based on my software development expierience (17 years) I can tell, that there is no clear borders between models of software development process organization. Every team is uniq. Therefore, there is no universal models. But there are rules:
- Rules to control access (based on ID, groups, roles, files, directories, branches, tags, schedules etc).
- Rules to replicate distributed repositories. Rules to roll-out new versions.
- Rules to resolve conflicts
- Even rules to create "smart" diffs
- and, of course, rules to "report" (if CVSWeb would be made with XSLT).
Prolog may help, but it's not a general programming language. Lisp is good, but it is not popular. Yet. Although, there is at least one project on Lisp related to CVS improvement: Meta-CVS. Python is much more popular and can be used effectively for AI programming, but there is no official standard for Python and that's why it is gonna dyi soon. Together with Perl. Erlang? Dylan? Too proprietary and there is not enough of libraries. The other languages are not good for functional programming. That's why I would support the effort of Meta-CVS, which is based on Lisp. Just I would go further and implement complete rule-based distributed change and configuration management layer over CVS.Then access control should be based on ACL (access control lists), rather than on primitive and unportable file attributes.
CVS already have triggers. Multi-site repository requires better access control and conflict resolution, probably both should be rule-based. But it sounds like workflow management, which is huge development - there is no good workflow management done open source, is it?
Should we boycott QT until then?
Correct me if I am wrong, but I thought about the future of 2.0 in terms of design and features of the version greater than 2.5 and, maybe, of 3.0. What is planned in future releases?
As for 2.0 itself - who cares about the dead meat. We must use 2.4 or 2.5. Period
It's not just linux - on any *n*x it's the same problem. The reason is the lack of something simular to NT domain and transparent network security.
More technically - OpenLDAP is supposed to solve it, but all vendors ignore it. All (almost) *n*x vendors still use obsolete (although often shadowed) password files in /etc instead of authentifying through OpenLDAp, which is is usually included to distro, but as something alike "contrib", for info, not for use. Without LDAP - no network-transparent authentication, network-transparent user-group security and thus - no secure way to share forlders.
Another general comment. *n*x came as a server - all users login from terminals to the same server where they share their folders under the same root "/". What is that "root" in network? Don't mention me already-depricated NFS - it's not for modern tasks. It's not reliable. And it's not secure.
Linux today is just coming to desktops (remember days of M$-Win 3.1?). Groupware environment is down the road. Be patient.
So lazy that don't use google to find anything about "linux X11 resolution", which point to lots of howtos and docs, including classic ones on www.linux.org and Linux Documentation Project.
Lazy novices should be among top ten reasons why good projects fail.
In Perl all you can do is moslty a primitive old-style regexp-oriented programming. Sort of grep or awk against XML files.
The reason is simple. Python as a programming language that has been designed by programmers. Perl is just a bad luck of its after-awk evolution done by programming system administrators.
- It works very stable for billions of years;
- It doesn't allow "bad" or useless information to exist too long;
- It's already known to modern human science;
Create some plants (I'm a vegetarian), better trees, encode all libraries to their DNA in order for information to be reproduced on tree leaves as pictures of texts with diagrams, better in English alphabytes, but Chineese or Cyrilic will work as well. Distribute such plants between public parks with a condition to grow such trees.Imagine - you go for BBQ and read some archived slashdot articles right on the tree :) Well, Slashdot might be wiped out by the evolution. But Einstein's articles will certainly stay forever! I am sure. Otherwise, why would the evolution allow Albert Einstein from the first place?
I enjoy the choice, I enjoy the development environment, I enjoy the ability to test my software in different GUI styles. I enjoy hte performance of my PostgreSQL, Zope, Tomcat, gcc compiler and various scriptings. I don't want buggy OS. I don't want stability and compatibility problems. I want the stability, compatibility and my freedom.
What about Mac OS X? It's the same crap as M$ winz. No future.
Three-Dimensional Workspace Manager, a Chalmers Medialab project, released as Free Software under the LGPL license and focusing on the development of a general-purpose platform for 3D user interfaces (3D GUIs). 3Dwm is NOT an X11 window manager and cannot work as such. It is a full 3D user environment, i.e. the 3D equivalent of X11 (far from completed, however).
Instead of confining your applications to the conventional 2D desktop, 3Dwm displays them in three dimensions and also provides the necessary for creating new applications with full 3D-UIs.
Designed especially for use in the Chalmers Medialab 3D-CUBE, 3Dwm will also run on most consumer-level desktop systems as well.
Install Cygwin on your winbox. You'll get a whole bunch of utilities from the real world helping you to feel yourself as not so isolated.
With Jabber on my Linux box both AIM and ICQ are interchangeable. If AOL will publicly ask for help I will do that. And of course a lot of other Jabber users can help :)
If you hate Java as unscalable proprietary lang then use either Python or CLOS/Lisp.
For those who argue about security of object attribute access - that has nothing to do with OOP and with security either. It's just a marketing blob and nothing useful. Hence, Python and CLOS/Lisp are still the choice.
Finally, CLOS/Lisp has not enough portable GUI libraries. Scheme is not OOP lang either. Thus we've left alone with Python. Portable open-source scalable stable-as-old and modern-as improved good OOP (and some FP!) lang Pyhton.
As for UNI*X features not yet available in Cygwin - you mean Gnome ported to win32, do you? At least that what failed the compilation Sawfish and others for me.
Have anyone heard about porting Gnome to win32? How about KDE? I can use Gnome or KDE through X11 net, but how about the case of temporary standalone box? Such win32 port would be a very nice feature then.
Open Source systems have a lot of alternatives. Big ones, like KDE vs GNOME, and small ones, like various PHP scripts. All this Open Source soup is evolving. And same as the mother-nature, it does it slow without sex (without exchange of source code pieces), or rapidly with sex (with exchange of source code pieces). Got a point? We have to steal software source code from each other in order to accelerate our software evolution!
- OS X
- Office X
- most of development and productivity software
It all costs money.Then ,there is a little of choice:
- I cannot change desktop manager
- I cannot change window manager
- How about the kernel?
All those themes change almost nothing - it's not a real customization I want as a user. And it's not a level of customization OEM would like either.Finally, it's not scalable down - it requires a lot of hardware resources and cannot work on low-end computer. It cannot be scaled up either - how about network application server? With both *n*x/X11 and M$/VNC it's easy.
Conclusion: Mac OS X is just a default OS installed on last Mac computers to make them more attractive within the sales process. But we all know that atractive-for-sales features are often different from useful-for-real-work features. Mac OS X is just a sales-toy. The more Linux is improved is the more often Linux will be installed after-sales.
Mac OS is the reason of low sales of Mac computers. Mac OS X is a pitty attempt to fix it. But only with modern Linux (pre-installed!)the situation with Mac sales can be improved.
When Steve Jobs will uderstand it, the production cost in Apple Corp wil be dropped and the sales will jump up.
I heard that Bill Gates payed ~$100M to Apple specially to keep them away from Linux. Isn't it true?
selling the stock, which you don't have - isn't it any different from BEC?
As for third... do we even really need any thirds?
Webmin is much better tool at least in its design. It reads real config files, so your manual changes will never be lost. It's still premature, though - some of modules, like usermin, are broken.
The major design problem of all such systems is a lack of versioning, (ACID) transactions and unified (XSLT?) reporting mechanism.
There is no pressure on that side from alternative vendors (Apple, microsoft), so I don't expext any big improvement here soon.
The best commercial admin env in Unix I saw was Smith in IBM AIX. It's object-orinted (sort of), good designed, and most important - it helps to learn CLI administration by giving the two-way access to CLI commands it runs. I wish IBM would contribute it to Linux at some day.
Wait. Apple has one unique feature - it's its hardware PPC platform. Wich works even better under Linux/PPC, which is more about Linux rather than about Apple.
And most of this market is either already troubled or it's gonna be troubled soon. Thanks to expensive Java.
Cost. The cost of developing and deploying a Java solution can be done with siginificantly less dollars ...
Platform elegance. Like it or leave it, the J2EE is a very nice development platform. Could it be better?
Yes, it could. Java is a statically typed language. Therefore, any dynamically typed OO language is more elegant - i.e. Python, Common Lisp.