I am incredibly lucky to live in a nice rural area, while it's lacking sometimes in some amenities it is, frankly, absolutely beautiful! Oh, and certainly much cheaper than trying to live in a city (which is another one of the reasons we ended up here).
Anyway, despite the estate being relatively new it appears that our internet connection is delivered via a damp piece of string and the maximum that ADSL 2 can give us is about 4.5Mb/sec. BT have started offering FTTC but even that only gets to 13Mb/sec according to the availability checker.
So we've had to go with keeping the rubbishy ADSL line for now (which is also on our landline) and getting Satellite Broadband as both a backup and for speed when needed - both of these really are to support my working/contracting from home where it's vital that I can connect to the office over VPN
Delivery
Actually arrived within a couple of days of the order! Big box with modem and TRIA and even bigger box with a white/cream Tooway satellite dish assembly. This is a BIG dish and rather heavy too. You could probably set it all up yourself but it's a fair weight to take up a height to fasten onto your wall so best to pay for someone to do that!
Fitting
Getting a professional fitting service meant that we had everything installed about a week after the order and it took about 2hrs to complete including running the cable to the sitting room and gettting it through the wall - same kind of procedure to having Sky TV installed.
Modem
The really important thing about the modem is that it is pretty basic. No configurable DHCP, no wireless, and no easy configuration to allow it to have a static IP address or play nice with the rest of your setup in anyway whatsoever. I can understand that the main 'cleverness' is in the satellite gadgetry but really the companies that make these things need some serious UX help
To insulate my carefully setup network from the modem I bought a TP-Link TL-R460 which is a neat little router that has a WAN port that can talk to modems over ethernet (it gets an IP address for this network via DHCP - yes, the SurfBeam2 Satellite Modem does have DHCP but it's not configurable) and then has a LAN network that 'looks' more like a normal router.
Usage
Primarily I'm using ADSL for now with the Satellite as backup. Eventually I'm going to use some sort of crude load-balancing or sharing of usage of the two links. The ADSL though slow is still quite responsive. The satellite has ping times around 700ms - 2000ms depending on site (I think the satellite ping time is around 600ms). That said it does still seem reasonably responsive.
I found the performance of the link to be extremely slow during the first week. The bandwidth was slower than the ADSL connection. After about a week of light usage/testing the performance started to get close to the promised 20Mb/sec download and 2Mb/sec upload. I need to carry out more tests though.
Overall
Definitely I need to make more use of the satellite to get a feel for performance and usefulness but for now it seems like it would be perfectly usable in an emergency and could also be handy for off-peak downloads/uploads. Despite the dish looking a bit of an eyesore it's at least hidden at the side of the house and isn't in the way.
I would also have to admit to a tiny glow of geekiness in that I'm using the internet over a satellite link! It's pretty cool that traffic can be sent direct from a dish on the house, up to a satellite, bounced back down to earth and then on to it's destination.
Showing posts with label geek. Show all posts
Showing posts with label geek. Show all posts
Tuesday, 2 December 2014
Friday, 6 January 2012
Consistent device links with udev
I have a Linux server that contains several disks used for holding photos, movies, videos and any other general junk myself and my partner dump on there. I've had some hassles with several of the drives and I also have one that is mounted in a caddy so it can be removed and hidden away (important data!).
What I found when things failed or needed to be moved was that having the drives mounted through fstab at boot time could lead to the machine just not starting. So I used automount to get round this problem. So far so good, but if one or more drives are removed the device names they use (/dev/sda /dev/sdb etc) change so my automount maps get screwed up.
After much searching I found some articles and bits and pieces that explained how to get around this issue using udev rules. I just had to re-fix this after an update to CentOS stopped these rules matching properly. Here is the file I use now:
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_ST31000524AS_9VPCT2B9", SYMLINK+="array0%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_ST31000524AS_9VPCS7ZT", SYMLINK+="array1%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_SAMSUNG_HD103SJ_S246J90B653757", SYMLINK+="array2%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_SAMSUNG_HD103SJ_S246J90B653776", SYMLINK+="array3%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_WDC_WD20EARS-00MVWB0_WD-WCAZA0998064", SYMLINK+="drives0%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_ST2000DL003-9VT166_6YD0EPBS", SYMLINK+="drives1%n"
The file is save as /etc/udev/rules.d/25-names.rules and looks for specific devices and creates links like /dev/array01 and /dev/drives11 depending on the ID it finds.
The bit I had to change was using %P for the parent device (/dev/sda1 has /dev/sda as a parent) and "sd*[1-9]" for the pattern to match only the partitions rather than the whole drive. Previously I used %k and a pattern that matched both the drive and partitions which did work before but not after the recent update.
What I found when things failed or needed to be moved was that having the drives mounted through fstab at boot time could lead to the machine just not starting. So I used automount to get round this problem. So far so good, but if one or more drives are removed the device names they use (/dev/sda /dev/sdb etc) change so my automount maps get screwed up.
After much searching I found some articles and bits and pieces that explained how to get around this issue using udev rules. I just had to re-fix this after an update to CentOS stopped these rules matching properly. Here is the file I use now:
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_ST31000524AS_9VPCT2B9", SYMLINK+="array0%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_ST31000524AS_9VPCS7ZT", SYMLINK+="array1%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_SAMSUNG_HD103SJ_S246J90B653757", SYMLINK+="array2%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_SAMSUNG_HD103SJ_S246J90B653776", SYMLINK+="array3%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_WDC_WD20EARS-00MVWB0_WD-WCAZA0998064", SYMLINK+="drives0%n"
KERNEL=="sd*[1-9]", BUS=="scsi", PROGRAM=="/sbin/scsi_id -g -u /dev/%P", RESULT=="1ATA_ST2000DL003-9VT166_6YD0EPBS", SYMLINK+="drives1%n"
The file is save as /etc/udev/rules.d/25-names.rules and looks for specific devices and creates links like /dev/array01 and /dev/drives11 depending on the ID it finds.
The bit I had to change was using %P for the parent device (/dev/sda1 has /dev/sda as a parent) and "sd*[1-9]" for the pattern to match only the partitions rather than the whole drive. Previously I used %k and a pattern that matched both the drive and partitions which did work before but not after the recent update.
Wednesday, 24 February 2010
Encoding Issues
A technical blog post for a change:
I've been having a really frustrating problem with my development environment recently, something that I've encountered before but never really understood or resolved definitively either: character encoding in source files.
Thanks to this blog post I think I've finally got a handle on what is going on and why. Essentially I had assumed that it was something odd in my environment but I think it's down to the fact that developers using Windows aren't setting the correct, UTF-8, encoding in their files which is therefore causing the problems for me.
The ultimate solution seems to be to fix these files by converting to UTF-8 and re-committing but I also have the issue that Maven appears to be using ASCII encoding by default so I'm not entirely sure if converting to UTF-8 will fix things. Unless of course it is expecting UTF-8 and is converting to ASCII to compile and the error message I'm getting is misleading. Will experiment later.
I've been having a really frustrating problem with my development environment recently, something that I've encountered before but never really understood or resolved definitively either: character encoding in source files.
Thanks to this blog post I think I've finally got a handle on what is going on and why. Essentially I had assumed that it was something odd in my environment but I think it's down to the fact that developers using Windows aren't setting the correct, UTF-8, encoding in their files which is therefore causing the problems for me.
The ultimate solution seems to be to fix these files by converting to UTF-8 and re-committing but I also have the issue that Maven appears to be using ASCII encoding by default so I'm not entirely sure if converting to UTF-8 will fix things. Unless of course it is expecting UTF-8 and is converting to ASCII to compile and the error message I'm getting is misleading. Will experiment later.
Wednesday, 4 November 2009
The Sky's the Limit
While alot of the house is still no longer visible I do remember that there are quite a few phone and television points so of course I wanted to get the PVR and TV plugged in and running so as not to miss all that wonderful winter programming (or probably more entertaining would be the Christmas adverts).
But something was wrong. Couldn't seem to get a signal at all from the main connection. Hmmm, I thought about this for a while and then happened to be in the back garden where I could see wires tacked to the wall - I assumed these were for the satellite dish ... and some were, but others were for the TV sockets ... and they were just connected together, no aerial in sight! It seems this house is limited to receiving programming from Sky!
So I need to look into the options: there is obviously Sky which is ridiculously over-priced, there is now FreeSat which seems better in that you simply need a receiver (though I's need to replace my PVR and some channels aren't available), I could also get an aerial fitted and stick to FreeView. All of these require money and complexity that probably aren't justified given the amount of TV I watch.
There is one other possibility on the horizon and that is SeeSaw. From the news report in the Guardian it seems that project Kangaroo has been resurrected and this could, if it aims to have the same content, offer an alternative to watching programmes over-the-air. It may also have the knock-on affect of increasing competition and driving down the costs of the likes of Sky and could maybe even lead to interesting home entertainment systems in the future.
For now, I think I'll just have to stick to iPlayer or my collection of DVDs for viewing entertainment, until I can save up to implement an alternative plan after some extensive research into the available gadgetry.
Friday, 9 October 2009
RAW processing on Linux
A technical post for a change (and the last entry for today as I've other things to be getting on with) and one that I hope will help others that may be struggling with the very same problem I faced.
I'd been using RawStudio on Ubuntu Linux for processing the RAW images from my Canon 350D camera. The results were fine and I didn't have any real problems, although I seemed to have to adjust the settings on almost every picture to get them to look right. I assumed that this was just down to what the camera captured and that having to tweak everything was just a natural thing to do.
I tried to use the same programme to handle images from Rachel's 500D camera and it got totally and utterly confused. The colours were completely off and while installing the latest daily snapshot did fix some of the issues it wasn't ideal. That said RawStudio is still a fantastic programme and it's very simple and easy to use.
Instead I found an application called RawTherapee which has proven to be much more successful and handles all the images flawlessly and, be default, appears to 'guess' the correct settings for all the images I've tried so far, both from the 350D and 500D cameras. The only criticism I have is that it's not as intuitive to use and is also complex in to deal with all of the available settings. However, all the available features means that it's incredibly powerful and I've already managed to 'rescue' some images because of it. For example there is a rotate feature which had meant some pictures that looked slightly wonky could be straightened very accurately and now look great.
To be completely fair, after a few sessions of use, I'm now getting use to the interface and I intend to use this for all the image processing from now on. I think I need to get my monitor calibrated at some point as well but that will have to wait until I can borrow something to achieve that (the tools are very expensive to buy apparently and the monitor I have - a Dell 2209WA (bought from Scan) - is meant to be fairly close to the right settings by default anyway.
Here's a screenshot of RawTherapee in action (in-expertly I should add, this is me using it here).
I'd been using RawStudio on Ubuntu Linux for processing the RAW images from my Canon 350D camera. The results were fine and I didn't have any real problems, although I seemed to have to adjust the settings on almost every picture to get them to look right. I assumed that this was just down to what the camera captured and that having to tweak everything was just a natural thing to do.
I tried to use the same programme to handle images from Rachel's 500D camera and it got totally and utterly confused. The colours were completely off and while installing the latest daily snapshot did fix some of the issues it wasn't ideal. That said RawStudio is still a fantastic programme and it's very simple and easy to use.
Instead I found an application called RawTherapee which has proven to be much more successful and handles all the images flawlessly and, be default, appears to 'guess' the correct settings for all the images I've tried so far, both from the 350D and 500D cameras. The only criticism I have is that it's not as intuitive to use and is also complex in to deal with all of the available settings. However, all the available features means that it's incredibly powerful and I've already managed to 'rescue' some images because of it. For example there is a rotate feature which had meant some pictures that looked slightly wonky could be straightened very accurately and now look great.
To be completely fair, after a few sessions of use, I'm now getting use to the interface and I intend to use this for all the image processing from now on. I think I need to get my monitor calibrated at some point as well but that will have to wait until I can borrow something to achieve that (the tools are very expensive to buy apparently and the monitor I have - a Dell 2209WA (bought from Scan) - is meant to be fairly close to the right settings by default anyway.
Here's a screenshot of RawTherapee in action (in-expertly I should add, this is me using it here).
Monday, 17 August 2009
Testing time
As part of my job hunting I had to take an online Java test which is meant to adapt to the answers you give and adjust the difficulty accordingly. Since I've been in permanent employment for the past four and a half years and before that not really having to do any serious job hunting, it's been a while since I've done anything like this.
So, how did I do? Well, I don't get to find out yet, the results are sent only to the company who wanted me to take the test. I did remember some of the questions and so I've checked my thoughts on these and I did catch at least one or two of the sneaky tricks they had, though equally there are a few I think they tripped me up on.
One of the big features of these tests is code that no competent person would right; it's legal according to the language specification (or at least parts of it are) but it would immediately get thrown out by even the most cursory peer review. This is why I always valued working as a trainer or mentoring, essentially those with less experience generally do 'think the unthinkable' and ask questions that you just never dreamed of!
Anyway, just have to wait and see how I did and hope it was enough.
Aside: I was looking into Java generics, trying to understand how it worked from the JCP book I have. The book spent pages and pages with examples and comments intended to put the reader at their ease by admitting that this was all a little confusing and didn't really work how you would expect.
Too be honest this rambling list of the differences, exceptions and important notes concerning rules you MUST remember for the exam was confusing me and I really just wanted to know how it all worked, how did all this functionality get represented at a low-level. So I ran javad and decompiled java.util.List to see how the magic worked.
What I discovered was that all the generics functionality is simply a compiler 'trick' and there is no under-lying implementation. The information that pertains to the generics syntax is stored in attributes which are an extensible feature of the classfile format. This article gives a pretty good description of the mechanism, which is admittedly quite neat.
Now why couldn't the book have explained all this? Armed with this knowledge all the complicated rules and exceptions make perfect sense! Why burden the reader with a multitude of disparate facts when knowing the how and why of something makes it so much clearer?!
Erm, rant over I think.
So, how did I do? Well, I don't get to find out yet, the results are sent only to the company who wanted me to take the test. I did remember some of the questions and so I've checked my thoughts on these and I did catch at least one or two of the sneaky tricks they had, though equally there are a few I think they tripped me up on.
One of the big features of these tests is code that no competent person would right; it's legal according to the language specification (or at least parts of it are) but it would immediately get thrown out by even the most cursory peer review. This is why I always valued working as a trainer or mentoring, essentially those with less experience generally do 'think the unthinkable' and ask questions that you just never dreamed of!
Anyway, just have to wait and see how I did and hope it was enough.
Aside: I was looking into Java generics, trying to understand how it worked from the JCP book I have. The book spent pages and pages with examples and comments intended to put the reader at their ease by admitting that this was all a little confusing and didn't really work how you would expect.
Too be honest this rambling list of the differences, exceptions and important notes concerning rules you MUST remember for the exam was confusing me and I really just wanted to know how it all worked, how did all this functionality get represented at a low-level. So I ran javad and decompiled java.util.List to see how the magic worked.
What I discovered was that all the generics functionality is simply a compiler 'trick' and there is no under-lying implementation. The information that pertains to the generics syntax is stored in attributes which are an extensible feature of the classfile format. This article gives a pretty good description of the mechanism, which is admittedly quite neat.
Now why couldn't the book have explained all this? Armed with this knowledge all the complicated rules and exceptions make perfect sense! Why burden the reader with a multitude of disparate facts when knowing the how and why of something makes it so much clearer?!
Erm, rant over I think.
Wednesday, 29 July 2009
Annoying GIMP!
Just spent a VERY frustrating 30mins trying desperately to make bezier curves using the path tool in the GIMP (Gnu Image Manipulation Programme), specifically I wanted to be able to turn the straight lines/vertices into more rounded shapes.
Normally you do this with 'handles' that are used to adjust the parameters for the curve, but despite following the instructions in the tutorials I just couldn't get these to appear!
After MUCH frustration and almost giving up I realised that I had the "Polygonal" option ticked (I was going to create a shape with many vertices and then curve these) which, according to the manual disables the handles and only allows straight lines! D'OH!
Normally you do this with 'handles' that are used to adjust the parameters for the curve, but despite following the instructions in the tutorials I just couldn't get these to appear!
After MUCH frustration and almost giving up I realised that I had the "Polygonal" option ticked (I was going to create a shape with many vertices and then curve these) which, according to the manual disables the handles and only allows straight lines! D'OH!
Wednesday, 3 June 2009
Jogger from memory lane
Okay, I'm sorry but this is a bit geeky.
Running ...
... failed ...
.... success .... YAY!
I'm sorry, I just thought it was kinda funny!
Just been testing something on an AIX server and I just happened to run the "smit" command which starts the admin tool, use to use this years and years ago when I did UNIX admin/development type stuff but haven't seen it since then. Anyway, I remembered the little running icon that it had from years ago, and guess what it's still there!
Running ...
... failed ...
.... success .... YAY!
I'm sorry, I just thought it was kinda funny!
Friday, 20 February 2009
London bound again
I'm on my way down to London again for another check-up appointment with Dr Curtis. This is the 3rd trip down and I guess I'm feeling pretty blase about it all since I'm so use to the journey and the experience of seeing a specialist.
There is probably more to discuss this time though and I think at this point the reality of what I'm doing is setting in and the hormones are, as I've mentioned, having mental as well as the physiological effects. I think a friend put it best when she said that I'll be feeling like I need to find myself, another also commented about the sense of a loss of control. I think the only other thought that I would add to this is the desire to be close to someone, a longing for the company of friends to talk to as well as someone special to be there to care about me and who I want to care for.
Hmm, there was an awful lot of flowery/poetic language in that last paragraph, will try not to make a habit of it!
Anyway, I'm actually sat on the train looking like a proper geek at the moment being as I'm connected to the internet via the 3G broadband dongle and I'm also doing some work (though this is rather slow going as it takes ages for the code I'm working on to compile, run, and then invariably fail requiring the whole process to be repeated!
But back to the mobile brodband: sometimes it's just nice to have some useful technology that just works, it makes the day just that little bit brighter :D
There is probably more to discuss this time though and I think at this point the reality of what I'm doing is setting in and the hormones are, as I've mentioned, having mental as well as the physiological effects. I think a friend put it best when she said that I'll be feeling like I need to find myself, another also commented about the sense of a loss of control. I think the only other thought that I would add to this is the desire to be close to someone, a longing for the company of friends to talk to as well as someone special to be there to care about me and who I want to care for.
Hmm, there was an awful lot of flowery/poetic language in that last paragraph, will try not to make a habit of it!
Anyway, I'm actually sat on the train looking like a proper geek at the moment being as I'm connected to the internet via the 3G broadband dongle and I'm also doing some work (though this is rather slow going as it takes ages for the code I'm working on to compile, run, and then invariably fail requiring the whole process to be repeated!
But back to the mobile brodband: sometimes it's just nice to have some useful technology that just works, it makes the day just that little bit brighter :D
Wednesday, 14 January 2009
Buzzing!
I know it is sad but I REALLY enjoy solving techie computer problems and getting things done! I'm happiest when there are loads of windows (xterms for preference) all running and doing useful things! Yes, very sad!
The problem with this level of nerdiness is that it kind of leaves your head a bit fuzzy and with your brain almost free-running and in desperate need of things to do and to figure out! It's a really bizarre feeling and almost like a drug because there is the downside when there is nothing to do and you almost get one of those downer-like feelings.
Anyway, I'm trying to counter-act the marathon typing/hacking/sorting-out session by drinking red wine, eating chicken, bacon and leek pie (more on this later) before going to bed! Need to get to bed early so I can be up first thing in the morning to get into work and get on with some serious coding!
Fizz, buzz, click, whir :D
The problem with this level of nerdiness is that it kind of leaves your head a bit fuzzy and with your brain almost free-running and in desperate need of things to do and to figure out! It's a really bizarre feeling and almost like a drug because there is the downside when there is nothing to do and you almost get one of those downer-like feelings.
Anyway, I'm trying to counter-act the marathon typing/hacking/sorting-out session by drinking red wine, eating chicken, bacon and leek pie (more on this later) before going to bed! Need to get to bed early so I can be up first thing in the morning to get into work and get on with some serious coding!
Fizz, buzz, click, whir :D
Sunday, 11 January 2009
Linux is better than Windows
Time for a geeky post I think! I remember commenting once that the transgendered aspect of my life was the thing that made me slightly more interesting that the rather geeky behaviour might infer. Now in some ways the trans thing is 'normal' and it's maybe time to let some of the extreme nerdiness out :)
There has been some talk at work over the last few days about setup of a server to host Subversion as a source code repository for the project that we are working on. Using Subversion was my idea and I pestered about it constantly last year for good technical reasons (it's alot quicker/easier to deal with that ClearCase which we'd used before) as well as the fact that I just like stuff like that!
And, when it came down to it the best platform to run Subversion on is Linux - I did try and be fair and install/run on Windows but when this was tested it was found that there were glitches under load which would have made the system to unreliable for what we needed.
Another thing I'm going to be looking at as well is running WebSphere without admin rights (found this article about allowing non-root users to create profiles). Why am I obsessed with admin rights - no-one at work seems to get this as the default setup of new laptops is to allow the user full admin rights which makes installing software easy.
This is just SOOOOOOOOOOoooooooo wrong if, like me you come from a UNIX/Linux background where running as root is a cardinal sin unless it's a single command prompt used for admin tasks that is closed immediately after finishing said tasks!
You see, even on Windows there is absolutely NO reason to have admin rights, there are no well-written programs that actually need such rights to work correctly. Okay, some services/servers need admin rights to run but they can either be started automatically or kicked off from a command prompt.
The problem is that with admin rights you can, quite easily move wreck your entire setup or at the very least move things around so that you've no real idea whether a program is installed properly or not. UNIX/Linux has been designed with this idea built in: important OS/program files can only be moved around by root and no-one but root can do this!
UNIX was designed to be multi-user from the start so the very idea that normal users would be able to muck around with important files would have been a huge issue since any one person could completely wreck everyone elses work! This level of security/protection is obviously essential for multi-user systems but has the side-benefit that if it's just you using the computer you know that nothing you do is going to affect anything important, all the libraries and other system components are going to behaving in a very perdictable and repeatable way.
And I think that is why, in my opinion, Linux is better than Windows; it's designed to be modular and consistent and essentially understandable, at least in part, at quite a deep level. UNIX/Linux doesn't generally suffer, in my experience, from bizarre interactions of libraries since everything is nicely protected and seperated. That's not to say things can't go wrong but when they do it's easier to isolate the problem and even work around it.
Also I prefer Linux because of the understandability (not sure that is a real word); I guess most people don't even bother trying to figure out how Windows works (I do but can't say I've got anywhere), but with Linux/UNIX there are some simple rules/concepts that, if you grasp them, then scale up for everything else within the system, e.g. command line tools for UNIX may seem complicated and harder to grasp than Windows DOS equivalent but the complexity is consistent in a way and once you get the idea of how everything can fit together it's simple.
Anyway, that is enough geekiness for now, I await the ridicule :)
There has been some talk at work over the last few days about setup of a server to host Subversion as a source code repository for the project that we are working on. Using Subversion was my idea and I pestered about it constantly last year for good technical reasons (it's alot quicker/easier to deal with that ClearCase which we'd used before) as well as the fact that I just like stuff like that!
And, when it came down to it the best platform to run Subversion on is Linux - I did try and be fair and install/run on Windows but when this was tested it was found that there were glitches under load which would have made the system to unreliable for what we needed.
Another thing I'm going to be looking at as well is running WebSphere without admin rights (found this article about allowing non-root users to create profiles). Why am I obsessed with admin rights - no-one at work seems to get this as the default setup of new laptops is to allow the user full admin rights which makes installing software easy.
This is just SOOOOOOOOOOoooooooo wrong if, like me you come from a UNIX/Linux background where running as root is a cardinal sin unless it's a single command prompt used for admin tasks that is closed immediately after finishing said tasks!
You see, even on Windows there is absolutely NO reason to have admin rights, there are no well-written programs that actually need such rights to work correctly. Okay, some services/servers need admin rights to run but they can either be started automatically or kicked off from a command prompt.
The problem is that with admin rights you can, quite easily move wreck your entire setup or at the very least move things around so that you've no real idea whether a program is installed properly or not. UNIX/Linux has been designed with this idea built in: important OS/program files can only be moved around by root and no-one but root can do this!
UNIX was designed to be multi-user from the start so the very idea that normal users would be able to muck around with important files would have been a huge issue since any one person could completely wreck everyone elses work! This level of security/protection is obviously essential for multi-user systems but has the side-benefit that if it's just you using the computer you know that nothing you do is going to affect anything important, all the libraries and other system components are going to behaving in a very perdictable and repeatable way.
And I think that is why, in my opinion, Linux is better than Windows; it's designed to be modular and consistent and essentially understandable, at least in part, at quite a deep level. UNIX/Linux doesn't generally suffer, in my experience, from bizarre interactions of libraries since everything is nicely protected and seperated. That's not to say things can't go wrong but when they do it's easier to isolate the problem and even work around it.
Also I prefer Linux because of the understandability (not sure that is a real word); I guess most people don't even bother trying to figure out how Windows works (I do but can't say I've got anywhere), but with Linux/UNIX there are some simple rules/concepts that, if you grasp them, then scale up for everything else within the system, e.g. command line tools for UNIX may seem complicated and harder to grasp than Windows DOS equivalent but the complexity is consistent in a way and once you get the idea of how everything can fit together it's simple.
Anyway, that is enough geekiness for now, I await the ridicule :)
Tuesday, 19 February 2008
Must ... not ... buy ... more gadgets
I've already bought one of those nice little Asus EEE PCs (and a fab little gadget it is too) but now there seems to be an even more affordable laptop soon to be released:
Elonex One Laptop
I've seen comments that this is in fact a re-badged version of Simple PC by Fontastic Telecom Inc but I don't have any inside information to confirm/deny this - they do appear to look awfully similar though!
The processor seems quite low-end, so I tried to have a hunt round for further specs. I managed to find the Aday Inc company website but the closest I could get to was the 5E CPU which is very interesting, being based on a 486 core!
So it's probably fair to assume that the Elonex One laptop could well be based on a high-end 486 CPU core (with all I/O on the same chip) or at best it's going to be a low-end Pentium CPU. Either way it just goes to show (assuming this new laptop actually does perform reasonably well) that you can still do an awful lot with fairly modest hardware!
Another point worth noting (just for a bit of Microsoft baiting) is that you are VERY unlikey to be able to get XP running on this machine, let alone Vista ... of course you may have a shot with something like Pocket PC/Windows Mobile/whatever it's called today - and in fairness the version of Linux on this device is likely to be extremely cut-down compared to a full distro - still, it appears to be running Pidgin which means it can handle GNOME and X-Windows.
Hmmm, I remember running a 486DX2-66 system with only 40Mb RAM and X-Windows, OLVWM, and being able to use Netscape Navigator, XEmacs, and other basic applications without any real issue .
Update: It looks as if this Aware A-Book AW-300 may be the same as the Elonex laptop. Flick through this review and there are some real revelations in terms of the OS and GUI for this device; it appears to be quite a mis-mash of applications and I even spotted some apps with Xt/CDE-style widgets rather than the more familar GTK/Qt that most modern OS's come with.
This is beginning to look like a real budget buy and nowhere near as useful, flexible, upgradable or generally polished as the Asus EEE PC or similar devices that are coming out. Then again, once hacked, a £99 device with screen, CPU, and plenty of I/O capabilities may have it's uses.
Elonex One Laptop
I've seen comments that this is in fact a re-badged version of Simple PC by Fontastic Telecom Inc but I don't have any inside information to confirm/deny this - they do appear to look awfully similar though!
The processor seems quite low-end, so I tried to have a hunt round for further specs. I managed to find the Aday Inc company website but the closest I could get to was the 5E CPU which is very interesting, being based on a 486 core!
So it's probably fair to assume that the Elonex One laptop could well be based on a high-end 486 CPU core (with all I/O on the same chip) or at best it's going to be a low-end Pentium CPU. Either way it just goes to show (assuming this new laptop actually does perform reasonably well) that you can still do an awful lot with fairly modest hardware!
Another point worth noting (just for a bit of Microsoft baiting) is that you are VERY unlikey to be able to get XP running on this machine, let alone Vista ... of course you may have a shot with something like Pocket PC/Windows Mobile/whatever it's called today - and in fairness the version of Linux on this device is likely to be extremely cut-down compared to a full distro - still, it appears to be running Pidgin which means it can handle GNOME and X-Windows.
Hmmm, I remember running a 486DX2-66 system with only 40Mb RAM and X-Windows, OLVWM, and being able to use Netscape Navigator, XEmacs, and other basic applications without any real issue .
Update: It looks as if this Aware A-Book AW-300 may be the same as the Elonex laptop. Flick through this review and there are some real revelations in terms of the OS and GUI for this device; it appears to be quite a mis-mash of applications and I even spotted some apps with Xt/CDE-style widgets rather than the more familar GTK/Qt that most modern OS's come with.
This is beginning to look like a real budget buy and nowhere near as useful, flexible, upgradable or generally polished as the Asus EEE PC or similar devices that are coming out. Then again, once hacked, a £99 device with screen, CPU, and plenty of I/O capabilities may have it's uses.
Friday, 8 February 2008
Is Linus irrelevant for Linux?
A brief break from the shopping and other TGirl pursuits; I was just reading this article:
The article seems confused; on one hand the author is saying that Linux is moving far ahead of Linus's limited vision and on the other he says, pointing to NetApp stats (which is meaningless really) that it is such a niche OS that it doesn't matter.
The star quote that is used as the main thrust of the authors argument is this from Linus
"An OS should never have been something that people (in general) really care about: it should be completely invisible and nobody should give a flying [expletive] about it except the technical people."
The author says that this "makes some sense" and then goes on to say that, in actual fact OS X and Windows have been successful because the customers DO care about the OS and thus Linus is no longer in touch with what the users want and is irrelevant.
The problem here is that the author clearly doesn't understand the subject he is talking about and is drawing conclusions from incorrect assumptions. The OS that he sees is all the flashy graphics and cool animations and features which is what the marketing types want you to see, this sells, this *is* relevant to end users.
In reality the *real* operating system is the stuff in the background that does all the boring work of, essentially, moving bits and bytes around the computer. This is what Linus means by the "operating system", not the marketing version of this term used to sell Vista, OS X and even Linux distros based on flashy features.
So the article is really trying to highlight an issue that isn't there; everyone knows that what users see as an "operating system" *is* important because it has been marketed to be, and yes Linus is not really involved with a great many aspects of what makes up such a product.
However, Linus still does have alot to do with the real OS and his opinion and voice on these aspects DOES matter.
Is Linus Torvalds even speaking for Linux anymore?
The article seems confused; on one hand the author is saying that Linux is moving far ahead of Linus's limited vision and on the other he says, pointing to NetApp stats (which is meaningless really) that it is such a niche OS that it doesn't matter.
The star quote that is used as the main thrust of the authors argument is this from Linus
"An OS should never have been something that people (in general) really care about: it should be completely invisible and nobody should give a flying [expletive] about it except the technical people."
The author says that this "makes some sense" and then goes on to say that, in actual fact OS X and Windows have been successful because the customers DO care about the OS and thus Linus is no longer in touch with what the users want and is irrelevant.
The problem here is that the author clearly doesn't understand the subject he is talking about and is drawing conclusions from incorrect assumptions. The OS that he sees is all the flashy graphics and cool animations and features which is what the marketing types want you to see, this sells, this *is* relevant to end users.
In reality the *real* operating system is the stuff in the background that does all the boring work of, essentially, moving bits and bytes around the computer. This is what Linus means by the "operating system", not the marketing version of this term used to sell Vista, OS X and even Linux distros based on flashy features.
So the article is really trying to highlight an issue that isn't there; everyone knows that what users see as an "operating system" *is* important because it has been marketed to be, and yes Linus is not really involved with a great many aspects of what makes up such a product.
However, Linus still does have alot to do with the real OS and his opinion and voice on these aspects DOES matter.
Subscribe to:
Posts (Atom)
