CPU and RAM are irrelevant if your applications are efficient. The Epoc/Nokia ones are and the Linux/Qtopia ones are not. Doesn't need 802.11b, it's a phone, wireless *everywhere*. GPRS, nice but I don't miss it.
You forgot to mention that the "advantage" of the open source software on the SL-6000 is that the user interface and applications suck badly, very very badly indeed and that the user interface and software on the Epoc 6 based Nokia doesn't.
Bluetooth is a lovely theory. Shame nothing is compatible in practice.
I have an SL-5500. I was suckered into buying one, but it's a dust gatherer. One of those gadgets like a toastie maker which you buy because it sounds like a good idea but turns out to be too much of a pain in the arse to use regularly. So they sit, gathering dust. The user interface and apps on the SL-6000 are virtually identical to the 5500, minor incremental differences so it'll also be a dust gatherer.
The Nokia on the other hand with a useful keyboard, a useful word processor, a useful agenda, a useful spreadsheet, a useful contacts manager, a useful route planner, a useful accounts package never leaves my side. Note the key word, *useful* unlike the dross installed on the SL-5500/6000 machines.
You get the lot in a single package. The best PDA OS and UI in the market, hundreds of third party applications and the phone is integrated right into the system. The cost sans connection is $699. If you get a connection with it, $50. Use the other $650 for something else.
I don't know what the problem is specifically, but if you have a couple of machines with KDE and Gnome loaded and a wireless (or other lower bandwidth higher latency) setup, you can demonstrate it yourself if you configure your display manager to allow remote logins for a while.
Gnome has a noticable lag or latency on virtually every operation; Menus, buttons, redraws. You might just rationalise it as network latency but...
KDE on the other hand is instantaneous for all of the above, it's as good as local. Same machine, same network connection and similar operations, the difference being the widget set. It's a very similar story for CDE and GnuStep. As far as I can see, Gnome is just slow.
Performance is vital. With X11 it isn't just the individual application performance which is important. The performance of the underlying widget set is also vitally important.
X11 is a network transparent display system. If the network performance of the GUI is unacceptable, it'll be virtually useless in a corporate environment. Unless you're an idiot and the best you can come up with is to put a full workstation on every desk.
For example. Running a full screen CDE session over an ADSL link is slow but perfectly usable. Gnome on the same machine over the same link on the other hand is completely unusable.
Preliminary (and subjective) testing indicates that it isn't good when compared to the competition; CDE, GnuStep and having just loaded the current KDE, it looks like that is faster as well. Testing commonly used stuff; Menu operations and such over a LAN.
Bugger... Anyone know of a platform I can build an objective test suite on, for the various competing GUIs rather than relying on a stopwatch? I've found lots of Java specific and Web specific stuff...
One of the benefits of Unix I suppose. On Windows you get Windows and so have nothing to compare with.
Re:Or... You could do it properly.
on
Optimizing distcc
·
· Score: 1
Maybe you need to send your sysadmin on the sysadmin course. It's free:
Our grid gets jobs out to an execution host and started in less than a second. All of our applications are distributed out over the execution nodes; Editors, word processors, spreadsheets, The Gimp, software builds, *everything*.
In fact, the less than 1 second latency incurred submitting a grid job is easily and by far overcome by the reduction in time given by starting a process on a machine which has an already running instance of that application. Netscape; sub 3 seconds, Open Office; sub 4 seconds. Emacs; only 50 or 60 seconds. You can achieve this by managing the queues.
We use the grid interactively for everything.
But then... I make a point of knowing what I'm doing.
This article is largely caused by the UK's overloaded transport infrastructure. Transport is a *problem*.
The irony is that the transport overload is caused by inane housing and planning policies.
Re:Or... You could do it properly.
on
Optimizing distcc
·
· Score: 1
Grid engine's a doddle to install and use. It's also more useful than a system limited to running distributed compiles.
Or... You could do it properly.
on
Optimizing distcc
·
· Score: 4, Informative
Install Sun Grid Engine[1] since it's free and now open source and then not only do you get qmake for distributed builds but you also get a general purpose distributed processing system. And hey! It even has the current buzzword "grid" in the title so your PHB will love you.
Just think about it as part of the cost of the box. The "build your own" numpties will have you spending hundreds more of your money and weeks of your time fiddling to get something which still doesn't do what tivo does within 10 minutes of plugging it in.
I mean, it is really really bad. All this reality crap? Shite game shows. Shite guest shows. Shite soap operas. My god they've found a lot of utter shite to spew out over hundreds of channels 24/7.
Tivo rocks though cos it finds the gems in the manure pile.
UML tools like Telelogic Tau go a fair way towards visual software development. There's a long way to go for the whole process to become driven by a completely visual interface though.
e.g. http://www.telelogic.com/products/tau/deve loper/in dex.cfm
It'd redefine the word bloat, that's for sure. Probably why Microsoft are interested. Bloat's what they do well.
No, the boxes I use are dual 3GHz machines
on
GNOME 2.6 Reviewed
·
· Score: 1
4Gb RAM, 15krpm SCSI disks, quad gigabit LAN and NO monitor.
They cost around 10k. Now. How many users will it support? 50? 100? 200? If you double the requirements you half the number of users and double the amount of money the corporation has to spend on new machines.
More important than that is the LAN performance. Have you got any idea how much it costs to put in gigabit switches and flood wire a building with cat5 or cat 6 cabling?
Why? Why? Why would you do this? Why not just put a machine on every desk? Because that's what idiots do. It's massively more expensive.
Absolutely nothing to do with hippies and everything to do with cold hard cash.
So how's performance? Or does it just not matter these days?
Look. Unlike Windows, this stuff is going to be going on to multi-user systems. There will be tens, hundreds of instances of each of the applications running on a particular machine... Over the network... Performance for X based applications is *absolutely crucial* in the corporate environment. That *is* where Gnome is going, isn't it?
Gnome 2.0 (Solaris packages) performs poorly in comparison to other X based UIs like CDE and Openstep. Both in local and network performance. So, does 2.6 suck or is it acceptable, is it even better?
Not with masses of plant and equipment. The costs of getting them there are pretty trivial, we already have plenty of probes on the planet. They just have to be able to carry an aerosol canister to disperse them. The hard part is designing microbes which will thrive and multiply in the environment.
Absolutely. We have already made that decision, billions of times. We do it every single day, every time you put a piece of meat in your mouth you make that decision.
Where do you draw the line? You draw the line with the greatest force. If they have the greater force you die and they live.
The human race will happily commit genocide in order that one set of genes will survive and another die out.
We are right here and right now wiping out species that compete in the food chain. Directly and through modification of their environment. Xenocide is a natural extension of this.
Slashdot Mirror
BTDT and not interested in dosemu.
I can get *serious* amounts of work done with WP5.1. Everything since has been downhill. So how well does the classic mode work on WP11?
CPU and RAM are irrelevant if your applications are efficient. The Epoc/Nokia ones are and the Linux/Qtopia ones are not. Doesn't need 802.11b, it's a phone, wireless *everywhere*. GPRS, nice but I don't miss it.
You forgot to mention that the "advantage" of the open source software on the SL-6000 is that the user interface and applications suck badly, very very badly indeed and that the user interface and software on the Epoc 6 based Nokia doesn't.
Bluetooth is a lovely theory. Shame nothing is compatible in practice.
I have an SL-5500. I was suckered into buying one, but it's a dust gatherer. One of those gadgets like a toastie maker which you buy because it sounds like a good idea but turns out to be too much of a pain in the arse to use regularly. So they sit, gathering dust. The user interface and apps on the SL-6000 are virtually identical to the 5500, minor incremental differences so it'll also be a dust gatherer.
The Nokia on the other hand with a useful keyboard, a useful word processor, a useful agenda, a useful spreadsheet, a useful contacts manager, a useful route planner, a useful accounts package never leaves my side. Note the key word, *useful* unlike the dross installed on the SL-5500/6000 machines.
You get the lot in a single package. The best PDA OS and UI in the market, hundreds of third party applications and the phone is integrated right into the system. The cost sans connection is $699. If you get a connection with it, $50. Use the other $650 for something else.
I don't know what the problem is specifically, but if you have a couple of machines with KDE and Gnome loaded and a wireless (or other lower bandwidth higher latency) setup, you can demonstrate it yourself if you configure your display manager to allow remote logins for a while.
Gnome has a noticable lag or latency on virtually every operation; Menus, buttons, redraws. You might just rationalise it as network latency but...
KDE on the other hand is instantaneous for all of the above, it's as good as local. Same machine, same network connection and similar operations, the difference being the widget set. It's a very similar story for CDE and GnuStep. As far as I can see, Gnome is just slow.
For anyone else interested in benchmarking X applications, there is a Perl module which is designed to perform user interaction.
http://sourceforge.net/projects/x11guitest/
It means that you can automate the testing of whole GUI applications or individual widgets.
"Performance is truly irrelevant."
What a stupid statement.
Performance is vital. With X11 it isn't just the individual application performance which is important. The performance of the underlying widget set is also vitally important.
X11 is a network transparent display system. If the network performance of the GUI is unacceptable, it'll be virtually useless in a corporate environment. Unless you're an idiot and the best you can come up with is to put a full workstation on every desk.
For example. Running a full screen CDE session over an ADSL link is slow but perfectly usable. Gnome on the same machine over the same link on the other hand is completely unusable.
The problem appears to be built into the widget set. Network performance is poor.
No mention?
Not important?
Preliminary (and subjective) testing indicates that it isn't good when compared to the competition; CDE, GnuStep and having just loaded the current KDE, it looks like that is faster as well. Testing commonly used stuff; Menu operations and such over a LAN.
Bugger... Anyone know of a platform I can build an objective test suite on, for the various competing GUIs rather than relying on a stopwatch? I've found lots of Java specific and Web specific stuff...
One of the benefits of Unix I suppose. On Windows you get Windows and so have nothing to compare with.
Maybe you need to send your sysadmin on the sysadmin course. It's free:
- 90 .html
http://suned.sun.com/US/catalog/courses/WE-1600
Our grid gets jobs out to an execution host and started in less than a second. All of our applications are distributed out over the execution nodes; Editors, word processors, spreadsheets, The Gimp, software builds, *everything*.
In fact, the less than 1 second latency incurred submitting a grid job is easily and by far overcome by the reduction in time given by starting a process on a machine which has an already running instance of that application. Netscape; sub 3 seconds, Open Office; sub 4 seconds. Emacs; only 50 or 60 seconds. You can achieve this by managing the queues.
We use the grid interactively for everything.
But then... I make a point of knowing what I'm doing.
And we have Linux all over the place. Replacing Windows and elderly Unix boxes at an accelerating pace.
We're ignoring SCO. Completely.
This article is largely caused by the UK's overloaded transport infrastructure. Transport is a *problem*.
The irony is that the transport overload is caused by inane housing and planning policies.
Grid engine's a doddle to install and use. It's also more useful than a system limited to running distributed compiles.
Install Sun Grid Engine[1] since it's free and now open source and then not only do you get qmake for distributed builds but you also get a general purpose distributed processing system. And hey! It even has the current buzzword "grid" in the title so your PHB will love you.
[1] http://gridengine.sunsource.net/
Yeah, I thought that too. Then I got Tivo and saw just how wrong I was.
Just think about it as part of the cost of the box. The "build your own" numpties will have you spending hundreds more of your money and weeks of your time fiddling to get something which still doesn't do what tivo does within 10 minutes of plugging it in.
I mean, it is really really bad. All this reality crap? Shite game shows. Shite guest shows. Shite soap operas. My god they've found a lot of utter shite to spew out over hundreds of channels 24/7.
Tivo rocks though cos it finds the gems in the manure pile.
UML tools like Telelogic Tau go a fair way towards visual software development. There's a long way to go for the whole process to become driven by a completely visual interface though.
e loper/in dex.cfm
e.g.
http://www.telelogic.com/products/tau/dev
It'd redefine the word bloat, that's for sure. Probably why Microsoft are interested. Bloat's what they do well.
4Gb RAM, 15krpm SCSI disks, quad gigabit LAN and NO monitor.
They cost around 10k. Now. How many users will it support? 50? 100? 200? If you double the requirements you half the number of users and double the amount of money the corporation has to spend on new machines.
More important than that is the LAN performance. Have you got any idea how much it costs to put in gigabit switches and flood wire a building with cat5 or cat 6 cabling?
Why? Why? Why would you do this? Why not just put a machine on every desk? Because that's what idiots do. It's massively more expensive.
Absolutely nothing to do with hippies and everything to do with cold hard cash.
So how's performance? Or does it just not matter these days?
Look. Unlike Windows, this stuff is going to be going on to multi-user systems. There will be tens, hundreds of instances of each of the applications running on a particular machine... Over the network... Performance for X based applications is *absolutely crucial* in the corporate environment. That *is* where Gnome is going, isn't it?
Gnome 2.0 (Solaris packages) performs poorly in comparison to other X based UIs like CDE and Openstep. Both in local and network performance. So, does 2.6 suck or is it acceptable, is it even better?
Can someone mirror it please?
"Infinite travel"? How do these things get modded insightful?
Not with masses of plant and equipment. The costs of getting them there are pretty trivial, we already have plenty of probes on the planet. They just have to be able to carry an aerosol canister to disperse them. The hard part is designing microbes which will thrive and multiply in the environment.
Absolutely. We have already made that decision, billions of times. We do it every single day, every time you put a piece of meat in your mouth you make that decision.
Where do you draw the line? You draw the line with the greatest force. If they have the greater force you die and they live.
The human race will happily commit genocide in order that one set of genes will survive and another die out.
We are right here and right now wiping out species that compete in the food chain. Directly and through modification of their environment. Xenocide is a natural extension of this.