Seaking about Solaris: it will be cheaper for Sun to switch Linux/Sparc as a primary low AND mid end OS. If they feel like Linux lacks some Solaris features - open those feature sources and people around the world wil port them to Linux in no time.
Speaking about Java - just admit that Java has failed. Java in general and EJB specifically are not scalable down, in terms of memory usage, process startup and small project development time. That is bad for really distributed applications. Admit also that a load-balanced cluster of small and mid end servers is cheaper and faster for 80% (if not 99%) of web (and many corporate) applications. In such situations the scalability up is also not really important. So, swtch to Python, Sun. And again, if you find that Python lacks some EJB (but not Java! - Python is practically perfect as a language) features - port them to Python, help Zope or 4Thought or Twisted projects.
In both cases switch your business model to consulting, customized solutions, training - learn from the success of IBM.
1. I use jabber.org from time to time - works stable and fine to me.
2. Most of time I use my own server, sharing it with few friends - never failed.
3. Whatever server you use - you don't have to use the same server as your friends as jabber namespaces will let your messages be fowarded and find the way to be delivered any way no matter if one of servers is down.
... then you may (should, must) appreciate Zope, especially its Plone portal implementation.
You may be interested alos in looking at an example of how and why the developer of formerly famous PHP-based forum has moved (re-wrote) the whole thing to Plone.
Here are some Zope successfull stories from the real market.
Well, then how about elimating a crappy programming (by making it impossible) in C++, like it's done since the begining in Java, Python, Smalltalk, Erlang, Lisp, Prolog, ML and Haskell.
"Internet Explorer" may not be cute, but by gosh nobody is going to be confused about what a product with that name is supposed to do. I'm really kind of flabbergasted that the Mozilla community can't come up with something, after months of discussion, that's better than Firebird/Thunderbird.
I like Chatzilla and other 'zillas. And I wonder why not rename Phoenix to Browzilla, while calling email program as Mailzilla? No one will have any doubts what program is doing as well as what technology it is based on.
One second = the time it takes for Cesium 133 to oscilate about 9.19 billion times (because it's something constant we can measure)
That's the reason of the problem! All other time constants are based on our astronomy, but the value of the second on Cesium. Even worth: about 9.19 billion times. "About"!
Certainly, the only way to solve the problem is to redefine the value of seconds. It must be based on the same astronomical measurements as other time values. Specifically, it must be 1/86400 of the day.
And in a digital era you don'thave to measure oscilation of any freaking Cesioum: use NTP instead!
Bad idea. It will be a penalty for successful ideas. Instead it must be a penalty for stupid ideas. So, the tax must be aka the property tax: $10,000 a year just for the right to keep a patent. If the idea is good then it will bring enough profit to pay such "patent lease fee". If not - you should have your right to reject your payent by yourself.
What is the economical reason to make SUB-orbital flights without any connections to futher ORBITAL programs? Just to show "Hey, mom! Look how high can I fly!", right?
As for launcing microsatellites, I doubt it's worthy to spend billions for rutan systems just b/c of that.
The best place would be Washington DC - that may clean the place from some garbage before another one may try find smarter solutions for the rest of the world.
By the way, US territory in general is a good place to clean up - the most weapons of mass destructios are there.
Compare the elegance of LISP brackets to overbloated XML tags. Compare the ability to share same syntax with code and data in LISP with oversimplified XML tags. Make a conclusion.
The time of procedural, no matter either structural or object oriented, languages is finished. With XML it is obvius that it was a mistake to ignore decarative languages.
Yeas, it was simpler to step from assembler to C to C++ and to Java. Because the compiler does really just a mapping from function calls to address jumps keeping the main idea of the program the same - do next step after you finished the previous one.
It is not the first time in our history when it was realized that the militarism is against the freedom and the spirit of software engineering. First experiments in 1960x, then Lisp, then Prolog, now XML/XSL.
Cycle by cycle, programmers required the language helping them to think in math about the problem rather than in electro-mechanical terms about how to control the computer behaviour.
And cycle by cycle the theory and practice of compilers (interpreters) failed to translate declarative description of the problem into the set of machine instructions. Did you noticed why? Because in case of procedural language it is a mapping from more structural and general set of instruction to concrete one specific to the machine. In case of declarative languages, it is not just a mapping, it is a real translation. The process is more complecated and the translator should be smart enough by itself. That time the world was not ready for smart translation. That's why most of declarative languages translated in interpreters. On more detail to notice: each declarative language had its own syntax in order to better reflect its semantics.
Time changes. With XML we've got a universal syntax capable to declare virtually any semantic. XML devides the programming progress into two big periods: without XML and with XML. Before XML the program language was based on "raw" text format and every compiler had own parser and own syntax validator. With XML the syntax parsing and validation problems become obsolete. With XML the compiler or rather interpreter will focus on semantic validation using formal human-readable language meta-description format such as Schema and RDF. Remember programming languages with exactly one line per instruction? Most of programmers even don't know such times. Well, in exception of legacy programmers still working with mainframes and hardware drivers. Same thing: people will forget non-XML syntax of programming languages. Again, in exception of legacy programmers.
Ok, I admit - there are just few XML-based programming languages today: XSL, XPath, XQuery and XUL are the most known. But now it just a metter of time when we will see all pre-XML programming paradigms in XML format: functional programming, backtracking, OOP. User Interface, workflow automation, data mining - every area will have own set of XML vocabularies and the real program will use ususally more than one vocabulary in one project. Why should I stick with one or two programming languages if I can create the right language for the right problem?
Did I fogot about Java and C#? Not yet. But I hope to forget at some point soon. Who remember Pascal and Algol? Same will be with Java and C#. You cannot write a hardware driver and expert system rules at the same language. That's why Java and C# have no future. C (actually, C++) has - we still need OS and GUI. We also need all those XML parsers and transformers. But why do I need JVM to do the task which is more alike a system software task? JVM is certainly not a very effective way to it. Any VM. Besides... XVM? What is XVM? eXtensible Virtual machine? Pure cross-platform AND cross-language universal LEGO-based engine? It is a concept we still have to understand.
I know you're out there. I can feel you now. I know that you're afraid... afraid of us. You're afraid of change. I don't know the future. I didn't come here to tell you how this is going to end. I came here to tell how it's going to begin. I'm going to hang up this phone, and then show these people what you don't want them to see. I'm going to show them a world without you. A world without rules or controls, borders or boundaries. A world where anything is possible. Where we go from there is a choice I leave to you.
or use AI in your daily projects. There are lots of area where AI could be used without being highlighted: DB rule-based queries, rule-based MOM, timeseria processing, adaptive workflow automation - you name it yourself.
All you need is some RDF to define desktop resources and XSL to translate RDF into any derived XML config files. Xalan will do the rest of XSLT for you. No wasting time on programming - do modelling instead.
go in IRC and try to find out who is the bot - it's already not so simple, somtimes. AI already not bad for textual based communications. And it would be improved soon drammatically with improvement in distributed fuzzy-indexing of text databases and with NLP in general. The improvement was not possible without breaking LAN boundaries - now we've got an internet. It was also not possible with the power of old PCs barely handling GUI - now the newest Apple dual G4 1Mhz has got a power way too much for GYU and web-surfing handling. We've got a vacuum of upcoming computing power, which is not consumed yet. Just hold a breath (just a decade) - and all that power will be consumed without any remains.
As for speech/voice recognition - once the speech would be translated to textual format then we come to the previos case. And some companies already work on it, like IBM. I think, we've got just few years (3 at least, 10 at most) of remained domination of keyboard/mouse interface. Touch-screens and voice scanners/commanders are coming up soon to change the face of all our PCs.
Mac OS X is not Mac OSv = 9.x, And Apple is just a couple of years with Mac OS X, various hardware support is just coming, be parient. Thus Apple's experience with Mac OS X is close to zero. A lot of intersting (and sad) surpises are just ahead.
The other point is in Apple's experience of Mac OS v = 9.x, which is the worst - bad system design (can you call that as a system at all?), very unstable, weak support of their own hardware platdform (heh, was it too hard to support the only platform?). M$'s quality is much better if you keep in mind that they support much broader base of hardware devices (how many competitors on the market of x86 motherboars and PCI/ISA cards?). And Linux was grown in the same "zoo" environment. Even more - Linux supports not only x86. I belive Linux even stronger in "zoo" support than BSD, especcially of non-server, mostly desktop-orineted.hardware, like firewire or scanning.
The reallity tells that Apple is good to bring RISC to the world of PCs - their hardware is worthy of the price. I can not tell it about their system software. Apple as a software company doesn't know how design/plan/control the quality of software. At least of system software.
At some point Apple realized the own failure and asked for the support from BSD community. Why BSD? Why not Linux? It's simple - they afraid Linux.as a competitor. BSD never directed to the desktop. But Linux did.
Finally, the choice between Mac OS X and Linux/PPC would driven more by the list of ported software, You need commercial packages ported for Mac OS X - boot Mac OS X. You need mostly Linux software - you welcome. However today i would not recommend to forget about stability. And with all other equal preferences today's choice choice is for Linux/PPC.
Slashdot Mirror
Speaking about Java - just admit that Java has failed. Java in general and EJB specifically are not scalable down, in terms of memory usage, process startup and small project development time. That is bad for really distributed applications. Admit also that a load-balanced cluster of small and mid end servers is cheaper and faster for 80% (if not 99%) of web (and many corporate) applications. In such situations the scalability up is also not really important. So, swtch to Python, Sun. And again, if you find that Python lacks some EJB (but not Java! - Python is practically perfect as a language) features - port them to Python, help Zope or 4Thought or Twisted projects.
In both cases switch your business model to consulting, customized solutions, training - learn from the success of IBM.
Or die.
that's in case of plaintext xdb. How about xdb_LDAP?
2. Most of time I use my own server, sharing it with few friends - never failed.
3. Whatever server you use - you don't have to use the same server as your friends as jabber namespaces will let your messages be fowarded and find the way to be delivered any way no matter if one of servers is down.
You may be interested alos in looking at an example of how and why the developer of formerly famous PHP-based forum has moved (re-wrote) the whole thing to Plone.
Here are some Zope successfull stories from the real market.
Apple Powerbook doesn't burn your jewels.
I wonder how many people on /. have computer WITHOUT ANY even a single piece made in China? 0%?
It's good to see one more nation in the space. Go China!
Remember, for every idiot-admin there is always some solution.
how is it different from Apple Powerbook based on PPC?
Well, then how about elimating a crappy programming (by making it impossible) in C++, like it's done since the begining in Java, Python, Smalltalk, Erlang, Lisp, Prolog, ML and Haskell.
I like Chatzilla and other 'zillas. And I wonder why not rename Phoenix to Browzilla, while calling email program as Mailzilla? No one will have any doubts what program is doing as well as what technology it is based on.
That's the reason of the problem! All other time constants are based on our astronomy, but the value of the second on Cesium. Even worth: about 9.19 billion times. "About"!
Certainly, the only way to solve the problem is to redefine the value of seconds. It must be based on the same astronomical measurements as other time values. Specifically, it must be 1/86400 of the day.
And in a digital era you don'thave to measure oscilation of any freaking Cesioum: use NTP instead!
Bad idea. It will be a penalty for successful ideas. Instead it must be a penalty for stupid ideas. So, the tax must be aka the property tax: $10,000 a year just for the right to keep a patent. If the idea is good then it will bring enough profit to pay such "patent lease fee". If not - you should have your right to reject your payent by yourself.
As for launcing microsatellites, I doubt it's worthy to spend billions for rutan systems just b/c of that.
By the way, US territory in general is a good place to clean up - the most weapons of mass destructios are there.
Compare the elegance of LISP brackets to overbloated XML tags. Compare the ability to share same syntax with code and data in LISP with oversimplified XML tags. Make a conclusion.
Yeas, it was simpler to step from assembler to C to C++ and to Java. Because the compiler does really just a mapping from function calls to address jumps keeping the main idea of the program the same - do next step after you finished the previous one.
It is not the first time in our history when it was realized that the militarism is against the freedom and the spirit of software engineering. First experiments in 1960x, then Lisp, then Prolog, now XML/XSL.
Cycle by cycle, programmers required the language helping them to think in math about the problem rather than in electro-mechanical terms about how to control the computer behaviour.
And cycle by cycle the theory and practice of compilers (interpreters) failed to translate declarative description of the problem into the set of machine instructions. Did you noticed why? Because in case of procedural language it is a mapping from more structural and general set of instruction to concrete one specific to the machine. In case of declarative languages, it is not just a mapping, it is a real translation. The process is more complecated and the translator should be smart enough by itself. That time the world was not ready for smart translation. That's why most of declarative languages translated in interpreters. On more detail to notice: each declarative language had its own syntax in order to better reflect its semantics.
Time changes. With XML we've got a universal syntax capable to declare virtually any semantic. XML devides the programming progress into two big periods: without XML and with XML. Before XML the program language was based on "raw" text format and every compiler had own parser and own syntax validator. With XML the syntax parsing and validation problems become obsolete. With XML the compiler or rather interpreter will focus on semantic validation using formal human-readable language meta-description format such as Schema and RDF. Remember programming languages with exactly one line per instruction? Most of programmers even don't know such times. Well, in exception of legacy programmers still working with mainframes and hardware drivers. Same thing: people will forget non-XML syntax of programming languages. Again, in exception of legacy programmers.
Ok, I admit - there are just few XML-based programming languages today: XSL, XPath, XQuery and XUL are the most known. But now it just a metter of time when we will see all pre-XML programming paradigms in XML format: functional programming, backtracking, OOP. User Interface, workflow automation, data mining - every area will have own set of XML vocabularies and the real program will use ususally more than one vocabulary in one project. Why should I stick with one or two programming languages if I can create the right language for the right problem?
Did I fogot about Java and C#? Not yet. But I hope to forget at some point soon. Who remember Pascal and Algol? Same will be with Java and C#. You cannot write a hardware driver and expert system rules at the same language. That's why Java and C# have no future. C (actually, C++) has - we still need OS and GUI. We also need all those XML parsers and transformers. But why do I need JVM to do the task which is more alike a system software task? JVM is certainly not a very effective way to it. Any VM. Besides... XVM? What is XVM? eXtensible Virtual machine? Pure cross-platform AND cross-language universal LEGO-based engine? It is a concept we still have to understand.
I know you're out there. I can feel you now. I know that you're afraid... afraid of us. You're afraid of change. I don't know the future. I didn't come here to tell you how this is going to end. I came here to tell how it's going to begin. I'm going to hang up this phone, and then show these people what you don't want them to see. I'm going to show them a world without you. A world without rules or controls, borders or boundaries. A world where anything is possible. Where we go from there is a choice I leave to you.
or use AI in your daily projects. There are lots of area where AI could be used without being highlighted: DB rule-based queries, rule-based MOM, timeseria processing, adaptive workflow automation - you name it yourself.
as money --> 0, knowledge --> infinity
we talked about XML as theme config files and I proposed to use a combination of RDF and XSL.
I am not against coding. But ... do you really suggest to code and compile theme config files?
... I don't think hackers love modeling - it would be for architects.
All you need is some RDF to define desktop resources and XSL to translate RDF into any derived XML config files. Xalan will do the rest of XSLT for you. No wasting time on programming - do modelling instead.
As for speech/voice recognition - once the speech would be translated to textual format then we come to the previos case. And some companies already work on it, like IBM. I think, we've got just few years (3 at least, 10 at most) of remained domination of keyboard/mouse interface. Touch-screens and voice scanners/commanders are coming up soon to change the face of all our PCs.
In fact, XUL would be a very good improvement for Glade/XML - if only GNOME's people would not so fail in love to M$.Net
The reason is in stability and support.
Mac OS X is not Mac OSv = 9.x, And Apple is just a couple of years with Mac OS X, various hardware support is just coming, be parient. Thus Apple's experience with Mac OS X is close to zero. A lot of intersting (and sad) surpises are just ahead.
The other point is in Apple's experience of Mac OS v = 9.x, which is the worst - bad system design (can you call that as a system at all?), very unstable, weak support of their own hardware platdform (heh, was it too hard to support the only platform?). M$'s quality is much better if you keep in mind that they support much broader base of hardware devices (how many competitors on the market of x86 motherboars and PCI/ISA cards?). And Linux was grown in the same "zoo" environment. Even more - Linux supports not only x86. I belive Linux even stronger in "zoo" support than BSD, especcially of non-server, mostly desktop-orineted.hardware, like firewire or scanning.
The reallity tells that Apple is good to bring RISC to the world of PCs - their hardware is worthy of the price. I can not tell it about their system software. Apple as a software company doesn't know how design/plan/control the quality of software. At least of system software.
At some point Apple realized the own failure and asked for the support from BSD community. Why BSD? Why not Linux? It's simple - they afraid Linux.as a competitor. BSD never directed to the desktop. But Linux did.
Finally, the choice between Mac OS X and Linux/PPC would driven more by the list of ported software, You need commercial packages ported for Mac OS X - boot Mac OS X. You need mostly Linux software - you welcome. However today i would not recommend to forget about stability. And with all other equal preferences today's choice choice is for Linux/PPC.
Can those boxes run Linux?