if you're not including the act of crafting the bits in the TCP stream by hand to assemble the HTTP requests, what's the point of even doing the comparison?
I think you're talking about Rails. Ruby as a web language can be very fast, especially with fcgi or using mongrel as the webserver.
Rails is "slow" because of all the automagic things it does, but it is possible to create a fast and responsive rails application even when it gets heavy users.
I wasn't switched to via fastuserswitching, but I do lock my screen. that seems to have an impact on it, too.
I ssh'd into my box at home and running this was successful.
fwiw, osascript doesn't work if the user isn't logged into aqua. I've tried writing volume controller scripts and I tried scripting Unison and other applications and they don't work if you're not logged in physically at the machine.
So basically, an exploit would need to be fired by the user or by something the user did (ie: surf to a website).
I just SSHed to my laptop and succesfully tested the above command.
I can't seem to get this to work. Not only does Applescript tell me that ARDAgent is not scriptable (when I tried to open it's scripting library:/System/Library/CoreServices/RemoteManagement/ARDAgent.app), but it also spit back this error:
ARDAgent got an error: "whoami" doesn't understand the do shell script message.
running the script on the commandline returns this:
spikedesktop:Library spike$ osascript -e 'tell app "ARDAgent" to do shell script "whoami"' 23:47: execution error: ARDAgent got an error: "whoami" doesnâ(TM)t understand the do shell script message. (-1708)
There have been numerous studies showing that SMART failure predictions are frequently incorrect, saying that a drive is not going to fail when it is, or is late in reporting a failure.
"most intriguing was that drives often needed replacement for issues that SMART drive status polling didn't or couldn't determine, and 56% of failed drives did not raise any significant SMART flags (and that's interesting, of course, because SMART exists solely to survey hard drive health)"
personally, I think SMART was developed by the drive companies to sell more drives. Frequently, admins (myself included) will replace a drive that they have a bad feeling about.
Drives are cheap enough that replacing a drive that possibly could fail is a trivial process.
do cars make people drive drunk? do purses make people thieves?
I understand what you're saying, but I think your metaphor is wrong.
A good example of what the article is saying is that people are more inclined to jump to google for an answer than do any real research. They're depending on google's accuracy for a solution.
Much like when you have a resident expert on PERL (for example), you will be more inclined to show them your code or ask them for help or input on something rather than doing the research or looking up the documentation on the library. By taking that shortcut, you're effectively stifling yourself and preventing your own skill growth.
People use Google as a crutch to support them rather than a vehicle to push their skills and excel.
This makes sure that their implementation is always the only one out there.
This is exactly how the patent system is broken. Patents on physical inventions leave room for competitors to invent their own version of the item, although the new inventor must create their own implementation that is sufficiently different. Patents on ideas and many business processes, especially this patent, are ridiculous. How can you patent a way of coming up with what software to right?
And how exactly is this different from SOA (Service Oriented Architecture) in the software world? I mean, you have one system that processes some set of data and then, after making it into a new pre-defined data structure (e.g. XML), it pushes it into another system which performs some other method of processing on it and then can return the results into a final service that takes that data and persists the information in a database.
The arbitrator/communicator element is essentially the primary orchestration element in SOA. The object/component driven element is essentially the portion that passes datastructures (XML) between systems. The situation/scenario driven element is just like the SOA as a whole when using choreography over orchestration.
Also, as other people have mentioned in other posts, it does feel like this is a patent on having levels of management. This is especially like the feel of a Graphics/PrePress shop where the salespeople interact with the clients, pass on job tickets to the manager of the prepress department who then checks over the tickets and distributes work to the worker bees.
Because the vast majority of people find such inferior quality completely acceptable.
yeah, that's really depressing.
Also depressing is how people accept DRM. My gf buys the majority of her music on iTunes and always says how the DRM doesn't affect her and that she doesn't mind. but it does affect her. I can't chuck her music on my ipod for my own personal listening (I don't want to grab her password even though she's offered it to me). Also, frequently she has issues because she's maxed out the number of computers she can authorize and doesn't want to de-authorize all machines because she thinks that'll mean she can't play it any more.
However, the quality of the vast majority of streaming music that's available online is not good enough for anyone who cares about what they listen to.
There are these services that are popping up left and right that enable you to download music from youtube (it basically rips the audio out of the FLV files and makes an mp3 that you can download or just creates a playlist of the video files without displaying the video for you to stream from your browser). I hate those things because the quality of youtube (both the video and the audio) are very low. It reminds me of what passed for normal desktop video in 98/99.
For the last decade I've been ripping my CDs the moment I get back to my computer and there are many tracks that I'd never listened to in their full quality. Being that I started ripping at 128kbps and switched to 192 shortly thereafter, I've been throwing out a big chunk of audio data. It wasn't until I listened to some full-quality, lossless tracks that I realized how much quality I was actually throwing away.
Low quality online-only audio is ok for streaming, especially if you're using it as background music from your PC speakers, but if you're going to listen on headphones or through any kind of decent speakers, even the iTunes purchased tracks aren't high quality enough... how can they expect us to pay [anything] for such inferior quality?
And how exactly are they going to determine that this mp3 is illegal and this other mp3 is not?
I've got a ton of music that I've ripped over the last 10 years using iTunes and nearly every other encoder as they became available over that time period (yes, I have to re-rip a lot because mpecker and other early encoders didn't do the hottest job). I've got a ton of music that I downloaded from the early days of eMusic when they had an all-you-can-eat service.
I've also got a ton of music that I downloaded because I lost the CDs or the CDs were too badly damaged to mp3 properly.
So, I'm curious what metric they are planning to use, exactly, to determine what content is "illegal." They better not deem that all non-iTunes purchased content is "illegal," just because it can be freely copied.
I've tried typing on the iphone with thumbs using the horizontal keyboard... I had a lot of issues with it; mostly they had to do with the fact that my thumbs obscure the keyboard which isn't as much the case with typing with the ol' index finger.
I understand what you're saying about muscle memory. I don't have an iPhone and I can totally understand that I would get used to it. When I first got a cellphone with a QWERTY keyboard (the original blackjack), I hated it. I was constantly mashing multiple keys, but the fact that the buttons were raised allowed me to more predictably press buttons that I couldn't see, specifically because I could feel them. Having feedback isn't as much of an issue as having in/outdentations for the buttons.
I've been touch-typing for over 15 years and have typed on a variety of keyboards including keyboards without nubs for my fingers and keyboards with nubs on weird keys... Sometimes they're on the index finger keys on home-row, sometimes the middle-finger... I used a keyboard that didn't have the little nubs once and it was confusing as hell, I had to keep looking down. I dabbled with that laser-projection keyboard thing last year and that was one of the most awkward typing experiences I've ever had. Because of lack of feedback (both tactile and knowing if I'm squarely hitting a button), I felt like I didn't improve my accuracy over time beyond the fact that I got a little used to it.
Learning how to type with thumbs only on a small device without any kind of physical indication as to edges on keys and with thumbs obscuring the view of the keypad... that doesn't sound fun. It's bad enough that my AT&T Tilt has a nearly perfectly smooth keyboard.
It would be sweet if an LCD could be manufactured that could dynamically adjust the texture of the surface to create nubs and key outlines and also give tactile feedback without obscuring the visual display too terribly.
I can't imagine using the iphone for any ssh access other than logging in, checking uptime, running a [existing] shell script or disabling a service. Having to do input with one finger, without any tactile feedback makes shell access sound like a chore.
I agree that the SDK may bring some new solutions, but seriously, I can't imagine being able to do anything truly productive in a terminal session with the iphone.
I do wish that they would just enable internet sharing on the iphone. I'd love to be able to use it as a wireless access point or to get online over bluetooth/usb on my laptop like I can with my AT&T Tilt.
gpg/pgp is great for the transfer... however once it's in the person's inbox, you have no idea what they're going to do with it.
Giving anyone other than my parents personal information about myself (credit card number/ SSN) over the phone pains me. It feels like I'm running a red light every time and I'd rather not do it.
Also handy for connecting to the serial console of other machines. When you can't ssh in to a machine because you screwed up the firewall, you can ssh into another machine on the network and accesss through the console.
Also handy for configuring managed switches (cisco, netgear, riverbed, juniper, etc).
Keyspan has the USB-serial thing covered, but needing to install drivers can get in the way sometimes.
How about you stop and think what specs PC's had at the beginning of the 90's, and still people somehow managed to get their stuff done.
I have this discussion with a lot of people. Do you want to use an OS from the beginning of the 90s?
Modern OSs and applications use more RAM and more resources than those from 10+ years ago. Between resiliency and robustness of data/filesystems/security (crypto, hashes, journals, metadata) and that modern websites contain 500%+ more code than sites did back then (javascript, CSS, xhtml, etc) and don't always function/render correctly without support for those features, you can't always use old-ass machines like that. If you're running applications like [Open]Office, the spellchecker and autosave features can consume all your resources on very low-end hardware (sub 200mhz). In windows, those machines will choke on antivirus and firewall software.
I've got a laptop (mentioned in another post) that's 300mhz w/ 128MB of RAM running Ubuntu with Blackbox for the windowmanager. It's almost unusable for most desktop tasks. It's great for ssh/vpn/rdp, but beyond that, firefox chokes and it can't handle a lot of other applications.
I'm not saying you need a multi-core machine with a gig of RAM... I'm just saying that you need something that's close-ish to 1ghz (6-800mhz should do) with a minimum of 256MB of RAM... RAM is cheap enough now that you should probably do 512MB-1GB. The eeePC would be perfect if there weren't much better machines coming out in the near future.
The only issue with older laptops is that they don't always get the greatest battery life... Modern subnotebooks (and hopefully some of the ones coming out in the next couple of months) should change that.
I've got an old Sony Vaio pcg-505tx. It's a Pentium1 @ 300mhz and 128MB of RAM. I picked up a Transcend 8GB IDE SSD for around $200 and popped that in there, too.
The downside is that it maxes out at 128MB of RAM, there's only 1 USB 1.0 port and it doesn't have onboard ethernet (but it does have PCMCIA). It doesn't allow booting over USB, so I had to use the [incredibly slow] CD drive that came with it which connects using the PCMCIA slot.
Just booting desktop linux eats up most of the memory in the box, so I disabled a lot of the system services. No sound, no automounter.
I changed gdm to a minimalist theme with no graphics and disabled gnome completely. I use Blackbox as my window manager. The upside is that once booted and logged into blackbox with no apps running, I've still got ~40MB of RAM free... launching a terminal window drops that down to about 24M of available RAM which is enough to ssh out to my server and check my email using mutt or to do edit code locally using vim.
I would not recommend this machine for running most modern applications. Firefox crushes it. I had to disable javascript because many websites cause the javascript VM to take so long, firefox complains and asks if you want to stop execution. Gmail doesn't work properly and you need to use the "light" or mobile versions. Apple's.mac email doesn't work.
I do have Aventail VPN working and I can RDP out using tsclient or rdesktop and that actually runs quite well, and I can get online using my cellphone (AT&T Tilt) using internet sharing over USB. I've found that it's actually faster to RDP over the VPN to work and surf the web than it is to do it locally.
The upside of the machine is that it gets ~3 hours of battery and it's incredibly light. It's much more practical to carry that around to check my email or read slashdot or VPN into work than it is to try to do that using my phone. I can barely feel that it's in my bag it's so light and small. Considering that I found the thing in the garbage a couple years ago (gotta love manhattan), it served as a great test to see if I'd really need/use a machine like that.
The result of the test? I love having a lightweight machine with me at all times. VPN and ssh and a full keyboard with a real OS everywhere I go is a godsend. I can't wait until this summer when machines start shipping with the atom processor. I'm picking up something to replace this vaio as soon as something decent comes out.
geeks.com have just jumped on this market. They have a couple of really nice lowish-end (in terms of expandability) 1U servers that would be great for such projects. I don't know if I'd use those in my datacenter, but they're great for home use.
I was looking at some of their dual-core xeon options with 2 SATA bays. You can get one of those for $400.
what features of VBA and Windows interoperability are most important to people.
Let me see.... all of them?
When I was writing VBA for Excel spreadsheets, I had this really great book on it (VBA for Excel), but many of the examples simply didn't work. This was mostly because of differences between VBA versions (I believe the windows version was 6 at the time, but the mac version was 4, but I could be way off... this is 4 years ago, I'm talking about).
This caused me to have extreme issues not only learning the language and techniques, but also in making sure a workbook that I created in windows would work in OSX and vice versa.
The vast majority of UI features were missing including many aspects of creating dropdown menus easily.
Also, the mac version ran several orders of magnitude slower:
I created a workbook that would import EZPass usage tables from the website and sort, find anomolies and colorize the output. Once the data was in the sheet, to do the aforementioned task was instantaneous in windows, but on the mac took over a minute. This is after I disabled window updates and implemented other performance enhancing tricks.
In the end, I wound up creating a webapp that would fetch the data, dump it to a file and have a perl script that would parse it and email a CSV file to be imported into excel for the boss. It didn't have the pretty colors (they wanted an excel file in the end, not just display in a browser), but it worked faster and required less user interaction.
so yeah, make the mac version of excel match the windows version. Afterall, isn't interoperability the point of having a multi-platform application?
Oh yeah, Microsoft's reason is to make money... and frustrate mac users into getting a windows box just to have a [more] pleasant experience (if you can call it that) with Office.
I've got an AT&T Tilt (HTC Tytan II) with HSDPA/3G/EDGE/GSM/etc and depending on where I am and what network, I get wildly different results. These is by using the bluetooth internet sharing with my MacbookPro in OSX or using USB internet sharing with my Ubuntu Linux Vaio:
Location / ping to google.com / max download speed At my dad's house in NJ / ~400-800ms / ~65K/sec NJTransit Train in NJ / ~80-90ms / ~110/sec NJX Airport / ~40-50ms / ~120K/sec In Brooklyn / ~70-80ms / ~120K/sec In Manhattan / ~40-50ms / ~120K/sec
I know when apple first switched to intel, everyone was happy to finally have an approximation as to what kinds of processors they'll see in the future since intel publishes their roadmaps and projections. Before, you had no idea what kind of chips they had in their labs (although, there were rumored 1ghz G3s before the G4 came out and 2ghz G4s before the G5 came out), and just because there were rumors, it didn't mean that those chips would ever see the light of day in any products.
I found an interesting add-on for linux and OSX a couple weeks ago (actually, it was posted in one of the linux mailing lists that I'm subscribed to, and I can't find the name of it at this very moment) that advertises itself as being a more robust AD client and allows your machines to take advantage of AD's centralized authentication and SSO facilities. I havne't actually installed it, but I downloaded it and am looking forward to checking it out. I feel a little wary about testing somethign like this out since I don't have an extra mac handy nor can I run OSX in a VM. I don't want to install it on any of the machines that I actually need to use. So, I can't vouch for it. But it looks promising since you don't need to run win2k3 R2 or update your current win2k3 schema to use it with a linux machine (OSX auths fine with a default win2k3 schema).
FYI, fast user logout sans QuickSilver is Shift-Opt-Cmd-Q. (you have to hold the keys about 1/3 second)
I'm not in front of my mac at the moment, but I'm pretty sure that that keyboard shortcut logs you out-out... like... quits all applications, etc. no? Or, actually... maybe that keyboard shortcut is cmd-shift-q. I really need to look into that, because it would be very exciting if you're right.
if it just brings up the login screen, that would be awesome.
And, regarding the keychain-screensaver option, I'll have to look into that. I wasn't aware that that was even there, but if it's just a screensaver, I dunno how much I like that; especially if it's not accessible from the keyboard.
Slashdot Mirror
if you're not including the act of crafting the bits in the TCP stream by hand to assemble the HTTP requests, what's the point of even doing the comparison?
I think you're talking about Rails. Ruby as a web language can be very fast, especially with fcgi or using mongrel as the webserver.
Rails is "slow" because of all the automagic things it does, but it is possible to create a fast and responsive rails application even when it gets heavy users.
Twitter is a perfect example of this.
Actually... interesting.
I wasn't switched to via fastuserswitching, but I do lock my screen. that seems to have an impact on it, too.
I ssh'd into my box at home and running this was successful.
fwiw, osascript doesn't work if the user isn't logged into aqua. I've tried writing volume controller scripts and I tried scripting Unison and other applications and they don't work if you're not logged in physically at the machine.
So basically, an exploit would need to be fired by the user or by something the user did (ie: surf to a website).
This is interesting.
I just SSHed to my laptop and succesfully tested the above command.
/System/Library/CoreServices/RemoteManagement/ARDAgent.app), but it also spit back this error:
I can't seem to get this to work. Not only does Applescript tell me that ARDAgent is not scriptable (when I tried to open it's scripting library:
ARDAgent got an error: "whoami" doesn't understand the do shell script message.
running the script on the commandline returns this:
spikedesktop:Library spike$ osascript -e 'tell app "ARDAgent" to do shell script "whoami"'
23:47: execution error: ARDAgent got an error: "whoami" doesnâ(TM)t understand the do shell script message. (-1708)
I'm running OSX 10.5.3 Intel.
SMART sucks
There have been numerous studies showing that SMART failure predictions are frequently incorrect, saying that a drive is not going to fail when it is, or is late in reporting a failure.
"most intriguing was that drives often needed replacement for issues that SMART drive status polling didn't or couldn't determine, and 56% of failed drives did not raise any significant SMART flags (and that's interesting, of course, because SMART exists solely to survey hard drive health)"
source:
http://www.engadget.com/2007/02/18/massive-google-hard-drive-survey-turns-up-very-interesting-thing/
personally, I think SMART was developed by the drive companies to sell more drives. Frequently, admins (myself included) will replace a drive that they have a bad feeling about.
Drives are cheap enough that replacing a drive that possibly could fail is a trivial process.
do cars make people drive drunk?
do purses make people thieves?
I understand what you're saying, but I think your metaphor is wrong.
A good example of what the article is saying is that people are more inclined to jump to google for an answer than do any real research. They're depending on google's accuracy for a solution.
Much like when you have a resident expert on PERL (for example), you will be more inclined to show them your code or ask them for help or input on something rather than doing the research or looking up the documentation on the library. By taking that shortcut, you're effectively stifling yourself and preventing your own skill growth.
People use Google as a crutch to support them rather than a vehicle to push their skills and excel.
This makes sure that their implementation is always the only one out there.
This is exactly how the patent system is broken. Patents on physical inventions leave room for competitors to invent their own version of the item, although the new inventor must create their own implementation that is sufficiently different. Patents on ideas and many business processes, especially this patent, are ridiculous. How can you patent a way of coming up with what software to right?
And how exactly is this different from SOA (Service Oriented Architecture) in the software world? I mean, you have one system that processes some set of data and then, after making it into a new pre-defined data structure (e.g. XML), it pushes it into another system which performs some other method of processing on it and then can return the results into a final service that takes that data and persists the information in a database.
The arbitrator/communicator element is essentially the primary orchestration element in SOA. The object/component driven element is essentially the portion that passes datastructures (XML) between systems. The situation/scenario driven element is just like the SOA as a whole when using choreography over orchestration.
Also, as other people have mentioned in other posts, it does feel like this is a patent on having levels of management. This is especially like the feel of a Graphics/PrePress shop where the salespeople interact with the clients, pass on job tickets to the manager of the prepress department who then checks over the tickets and distributes work to the worker bees.
yeah, everyone knows that. ;)
I've also purchased a license for PodWorks.
but that's not my problem.... I still can't play the iTMS files.
Because the vast majority of people find such inferior quality completely acceptable.
yeah, that's really depressing.
Also depressing is how people accept DRM. My gf buys the majority of her music on iTunes and always says how the DRM doesn't affect her and that she doesn't mind. but it does affect her. I can't chuck her music on my ipod for my own personal listening (I don't want to grab her password even though she's offered it to me). Also, frequently she has issues because she's maxed out the number of computers she can authorize and doesn't want to de-authorize all machines because she thinks that'll mean she can't play it any more.
How many people really want music that can only be played from the internet? For some people this would work, sure.
And take into account that they are putting up much lower quality tracks to reduce filesize to save on bandwidth costs. Hardly worth it.
However, the quality of the vast majority of streaming music that's available online is not good enough for anyone who cares about what they listen to.
There are these services that are popping up left and right that enable you to download music from youtube (it basically rips the audio out of the FLV files and makes an mp3 that you can download or just creates a playlist of the video files without displaying the video for you to stream from your browser). I hate those things because the quality of youtube (both the video and the audio) are very low. It reminds me of what passed for normal desktop video in 98/99.
For the last decade I've been ripping my CDs the moment I get back to my computer and there are many tracks that I'd never listened to in their full quality. Being that I started ripping at 128kbps and switched to 192 shortly thereafter, I've been throwing out a big chunk of audio data. It wasn't until I listened to some full-quality, lossless tracks that I realized how much quality I was actually throwing away.
Low quality online-only audio is ok for streaming, especially if you're using it as background music from your PC speakers, but if you're going to listen on headphones or through any kind of decent speakers, even the iTunes purchased tracks aren't high quality enough... how can they expect us to pay [anything] for such inferior quality?
And how exactly are they going to determine that this mp3 is illegal and this other mp3 is not?
I've got a ton of music that I've ripped over the last 10 years using iTunes and nearly every other encoder as they became available over that time period (yes, I have to re-rip a lot because mpecker and other early encoders didn't do the hottest job). I've got a ton of music that I downloaded from the early days of eMusic when they had an all-you-can-eat service.
I've also got a ton of music that I downloaded because I lost the CDs or the CDs were too badly damaged to mp3 properly.
So, I'm curious what metric they are planning to use, exactly, to determine what content is "illegal." They better not deem that all non-iTunes purchased content is "illegal," just because it can be freely copied.
I've tried typing on the iphone with thumbs using the horizontal keyboard... I had a lot of issues with it; mostly they had to do with the fact that my thumbs obscure the keyboard which isn't as much the case with typing with the ol' index finger.
I understand what you're saying about muscle memory. I don't have an iPhone and I can totally understand that I would get used to it. When I first got a cellphone with a QWERTY keyboard (the original blackjack), I hated it. I was constantly mashing multiple keys, but the fact that the buttons were raised allowed me to more predictably press buttons that I couldn't see, specifically because I could feel them. Having feedback isn't as much of an issue as having in/outdentations for the buttons.
I've been touch-typing for over 15 years and have typed on a variety of keyboards including keyboards without nubs for my fingers and keyboards with nubs on weird keys... Sometimes they're on the index finger keys on home-row, sometimes the middle-finger... I used a keyboard that didn't have the little nubs once and it was confusing as hell, I had to keep looking down. I dabbled with that laser-projection keyboard thing last year and that was one of the most awkward typing experiences I've ever had. Because of lack of feedback (both tactile and knowing if I'm squarely hitting a button), I felt like I didn't improve my accuracy over time beyond the fact that I got a little used to it.
Learning how to type with thumbs only on a small device without any kind of physical indication as to edges on keys and with thumbs obscuring the view of the keypad... that doesn't sound fun. It's bad enough that my AT&T Tilt has a nearly perfectly smooth keyboard.
It would be sweet if an LCD could be manufactured that could dynamically adjust the texture of the surface to create nubs and key outlines and also give tactile feedback without obscuring the visual display too terribly.
I can't imagine using the iphone for any ssh access other than logging in, checking uptime, running a [existing] shell script or disabling a service. Having to do input with one finger, without any tactile feedback makes shell access sound like a chore.
I agree that the SDK may bring some new solutions, but seriously, I can't imagine being able to do anything truly productive in a terminal session with the iphone.
I do wish that they would just enable internet sharing on the iphone. I'd love to be able to use it as a wireless access point or to get online over bluetooth/usb on my laptop like I can with my AT&T Tilt.
gpg/pgp is great for the transfer... however once it's in the person's inbox, you have no idea what they're going to do with it.
Giving anyone other than my parents personal information about myself (credit card number/ SSN) over the phone pains me. It feels like I'm running a red light every time and I'd rather not do it.
Also handy for connecting to the serial console of other machines. When you can't ssh in to a machine because you screwed up the firewall, you can ssh into another machine on the network and accesss through the console.
Also handy for configuring managed switches (cisco, netgear, riverbed, juniper, etc).
Keyspan has the USB-serial thing covered, but needing to install drivers can get in the way sometimes.
How about you stop and think what specs PC's had at the beginning of the 90's, and still people somehow managed to get their stuff done.
I have this discussion with a lot of people. Do you want to use an OS from the beginning of the 90s?
Modern OSs and applications use more RAM and more resources than those from 10+ years ago. Between resiliency and robustness of data/filesystems/security (crypto, hashes, journals, metadata) and that modern websites contain 500%+ more code than sites did back then (javascript, CSS, xhtml, etc) and don't always function/render correctly without support for those features, you can't always use old-ass machines like that. If you're running applications like [Open]Office, the spellchecker and autosave features can consume all your resources on very low-end hardware (sub 200mhz). In windows, those machines will choke on antivirus and firewall software.
I've got a laptop (mentioned in another post) that's 300mhz w/ 128MB of RAM running Ubuntu with Blackbox for the windowmanager. It's almost unusable for most desktop tasks. It's great for ssh/vpn/rdp, but beyond that, firefox chokes and it can't handle a lot of other applications.
I'm not saying you need a multi-core machine with a gig of RAM... I'm just saying that you need something that's close-ish to 1ghz (6-800mhz should do) with a minimum of 256MB of RAM... RAM is cheap enough now that you should probably do 512MB-1GB. The eeePC would be perfect if there weren't much better machines coming out in the near future.
The only issue with older laptops is that they don't always get the greatest battery life... Modern subnotebooks (and hopefully some of the ones coming out in the next couple of months) should change that.
I've got an old Sony Vaio pcg-505tx. It's a Pentium1 @ 300mhz and 128MB of RAM. I picked up a Transcend 8GB IDE SSD for around $200 and popped that in there, too.
.mac email doesn't work.
The downside is that it maxes out at 128MB of RAM, there's only 1 USB 1.0 port and it doesn't have onboard ethernet (but it does have PCMCIA). It doesn't allow booting over USB, so I had to use the [incredibly slow] CD drive that came with it which connects using the PCMCIA slot.
Just booting desktop linux eats up most of the memory in the box, so I disabled a lot of the system services. No sound, no automounter.
I changed gdm to a minimalist theme with no graphics and disabled gnome completely. I use Blackbox as my window manager. The upside is that once booted and logged into blackbox with no apps running, I've still got ~40MB of RAM free... launching a terminal window drops that down to about 24M of available RAM which is enough to ssh out to my server and check my email using mutt or to do edit code locally using vim.
I would not recommend this machine for running most modern applications. Firefox crushes it. I had to disable javascript because many websites cause the javascript VM to take so long, firefox complains and asks if you want to stop execution. Gmail doesn't work properly and you need to use the "light" or mobile versions. Apple's
I do have Aventail VPN working and I can RDP out using tsclient or rdesktop and that actually runs quite well, and I can get online using my cellphone (AT&T Tilt) using internet sharing over USB. I've found that it's actually faster to RDP over the VPN to work and surf the web than it is to do it locally.
The upside of the machine is that it gets ~3 hours of battery and it's incredibly light. It's much more practical to carry that around to check my email or read slashdot or VPN into work than it is to try to do that using my phone. I can barely feel that it's in my bag it's so light and small. Considering that I found the thing in the garbage a couple years ago (gotta love manhattan), it served as a great test to see if I'd really need/use a machine like that.
The result of the test? I love having a lightweight machine with me at all times. VPN and ssh and a full keyboard with a real OS everywhere I go is a godsend. I can't wait until this summer when machines start shipping with the atom processor. I'm picking up something to replace this vaio as soon as something decent comes out.
geeks.com have just jumped on this market. They have a couple of really nice lowish-end (in terms of expandability) 1U servers that would be great for such projects. I don't know if I'd use those in my datacenter, but they're great for home use.
I was looking at some of their dual-core xeon options with 2 SATA bays. You can get one of those for $400.
erm... yeah I meant EWR...
I think I accidently typed the name of the workgroup for one of my old employers... hehe.
I knew 'NJX' didn't look right.
what features of VBA and Windows interoperability are most important to people.
Let me see.... all of them?
When I was writing VBA for Excel spreadsheets, I had this really great book on it (VBA for Excel), but many of the examples simply didn't work. This was mostly because of differences between VBA versions (I believe the windows version was 6 at the time, but the mac version was 4, but I could be way off... this is 4 years ago, I'm talking about).
This caused me to have extreme issues not only learning the language and techniques, but also in making sure a workbook that I created in windows would work in OSX and vice versa.
The vast majority of UI features were missing including many aspects of creating dropdown menus easily.
Also, the mac version ran several orders of magnitude slower:
I created a workbook that would import EZPass usage tables from the website and sort, find anomolies and colorize the output. Once the data was in the sheet, to do the aforementioned task was instantaneous in windows, but on the mac took over a minute. This is after I disabled window updates and implemented other performance enhancing tricks.
In the end, I wound up creating a webapp that would fetch the data, dump it to a file and have a perl script that would parse it and email a CSV file to be imported into excel for the boss. It didn't have the pretty colors (they wanted an excel file in the end, not just display in a browser), but it worked faster and required less user interaction.
so yeah, make the mac version of excel match the windows version. Afterall, isn't interoperability the point of having a multi-platform application?
Oh yeah, Microsoft's reason is to make money... and frustrate mac users into getting a windows box just to have a [more] pleasant experience (if you can call it that) with Office.
I've got an AT&T Tilt (HTC Tytan II) with HSDPA/3G/EDGE/GSM/etc and depending on where I am and what network, I get wildly different results. These is by using the bluetooth internet sharing with my MacbookPro in OSX or using USB internet sharing with my Ubuntu Linux Vaio:
Location / ping to google.com / max download speed
At my dad's house in NJ / ~400-800ms / ~65K/sec
NJTransit Train in NJ / ~80-90ms / ~110/sec
NJX Airport / ~40-50ms / ~120K/sec
In Brooklyn / ~70-80ms / ~120K/sec
In Manhattan / ~40-50ms / ~120K/sec
I want to get a tattoo of myself on my entire body, only 2" taller.
No roadmap.
Yeah, another great point.
I know when apple first switched to intel, everyone was happy to finally have an approximation as to what kinds of processors they'll see in the future since intel publishes their roadmaps and projections. Before, you had no idea what kind of chips they had in their labs (although, there were rumored 1ghz G3s before the G4 came out and 2ghz G4s before the G5 came out), and just because there were rumors, it didn't mean that those chips would ever see the light of day in any products.
I found an interesting add-on for linux and OSX a couple weeks ago (actually, it was posted in one of the linux mailing lists that I'm subscribed to, and I can't find the name of it at this very moment) that advertises itself as being a more robust AD client and allows your machines to take advantage of AD's centralized authentication and SSO facilities. I havne't actually installed it, but I downloaded it and am looking forward to checking it out. I feel a little wary about testing somethign like this out since I don't have an extra mac handy nor can I run OSX in a VM. I don't want to install it on any of the machines that I actually need to use. So, I can't vouch for it. But it looks promising since you don't need to run win2k3 R2 or update your current win2k3 schema to use it with a linux machine (OSX auths fine with a default win2k3 schema).
FYI, fast user logout sans QuickSilver is Shift-Opt-Cmd-Q. (you have to hold the keys about 1/3 second)
I'm not in front of my mac at the moment, but I'm pretty sure that that keyboard shortcut logs you out-out... like... quits all applications, etc. no? Or, actually... maybe that keyboard shortcut is cmd-shift-q. I really need to look into that, because it would be very exciting if you're right.
if it just brings up the login screen, that would be awesome.
And, regarding the keychain-screensaver option, I'll have to look into that. I wasn't aware that that was even there, but if it's just a screensaver, I dunno how much I like that; especially if it's not accessible from the keyboard.