Friday 17 October 2008, 3:51 AM
Marketing Mush
I felt like I was wading through mush reading the Engineering Windows 7 blog. I've waited until now to read all of the accumulated postings. Little to no details except what they are "focusing" on while they do whatever they are doing to bring Windows 7 to market. The blogs read more like techie versions of marketing focus-group reports. "Mojave" for the technically inclined maybe.
I was a little dismayed to see that they think the best way to use Windows (any version back to and including XP) is to have the system come up from "sleep" mode. Yes that makes it boot faster more often, but it also means you have to leave the system powered while doing that. Yes its at a much lower power setting BUT it still requires running current into the system. I would like them come up with a method to have the system come back up from a cold dead stop in 5 seconds or less. Its going to require a combination of hardware and software but it wouldn't be the first time Microsoft dictated terms to the hardware vendors. Windows 95 "compatible" CRT monitors comes immediately to mind.
If Radio Shack could do it more than 15 years ago with the CMOS TRS80 model 2, why can't it be done now with the vastly superior laptop and netbook technology of today?
I was also dismayed that most of the discussion seemed to be centered on Vista and what they wanted to do with it. Their focus in the later blogs seemed to hinge on justifications for what was in Vista and how they were going to "fix" the settings in Win7. hmmm. They make it sound more like Vista SP2 or Vista2 or maybe they ought to call it The Mojave Vista.
However, taking the blogs at face value, I'm glad that they seem to be interested in making the slow-as-molasses processes in Vista run faster in Win7. Whoopee. Installing hardware drivers in "parallel" on boot-up is something that is easily done in Linux. Just change a setting in Debian and it boots nearly twice as fast as it used to. Currently DebbieToo (a Celeron) and SugarBear (a P3) both boot up faster than Windows XP SP3 on my P4.
I've been reading a couple of books about embedded Linux. It looks like there are a number of other things I can do to speed up boot time and application run times for Debian or other Linux versions. I also will be able to customize the operating system image to only include what I want in it and not what the Redmond Gorilla wants in it. I doubt that I'll have that sort of flexibility to speed up or customize Windows 7, or whatever the marketing hacks decide to call it!
I was a little dismayed to see that they think the best way to use Windows (any version back to and including XP) is to have the system come up from "sleep" mode. Yes that makes it boot faster more often, but it also means you have to leave the system powered while doing that. Yes its at a much lower power setting BUT it still requires running current into the system. I would like them come up with a method to have the system come back up from a cold dead stop in 5 seconds or less. Its going to require a combination of hardware and software but it wouldn't be the first time Microsoft dictated terms to the hardware vendors. Windows 95 "compatible" CRT monitors comes immediately to mind.
If Radio Shack could do it more than 15 years ago with the CMOS TRS80 model 2, why can't it be done now with the vastly superior laptop and netbook technology of today?
I was also dismayed that most of the discussion seemed to be centered on Vista and what they wanted to do with it. Their focus in the later blogs seemed to hinge on justifications for what was in Vista and how they were going to "fix" the settings in Win7. hmmm. They make it sound more like Vista SP2 or Vista2 or maybe they ought to call it The Mojave Vista.
However, taking the blogs at face value, I'm glad that they seem to be interested in making the slow-as-molasses processes in Vista run faster in Win7. Whoopee. Installing hardware drivers in "parallel" on boot-up is something that is easily done in Linux. Just change a setting in Debian and it boots nearly twice as fast as it used to. Currently DebbieToo (a Celeron) and SugarBear (a P3) both boot up faster than Windows XP SP3 on my P4.
I've been reading a couple of books about embedded Linux. It looks like there are a number of other things I can do to speed up boot time and application run times for Debian or other Linux versions. I also will be able to customize the operating system image to only include what I want in it and not what the Redmond Gorilla wants in it. I doubt that I'll have that sort of flexibility to speed up or customize Windows 7, or whatever the marketing hacks decide to call it!
Sunday 12 October 2008, 8:57 PM
When isn't a File Write Permission? When its SP3?
I've been working a lot lately on a project that as part of its functionality writes over or deletes files in a sub-folder in the Program Files folder in Windows XP Pro. Trying to delete the files or overwrite them using a VB.Net program written using VS2005/SP1 throws an exception error. The program has a local signed certificate in a totally trusted mode. If the files don't exist, the program can easily write them as new files without an exception
The user logon at both times (during the previous “rename-the-computer” boot and the “current” boot-up) are the same logged-on user who's an Administrator group member. Just the local “domain” name, the computer name has changed. The users have remained the same, the file permissions and the group permissions have remained the same etc.
In all cases, the computer has remained in Workgroup mode. There has been no attempt, at any time, to join a NT or Active Directory domain performed on the computer OS image load.
DEP is on only for Windows OS files.
The OS image starts out as a ghost image, that was a SP3 slipstreamed volume license of Windows XP Pro SP1. In other words, its running as a Windows Pro SP3 system. It operates just fine, no errors, no BSODs etc. You've already heard about the Router-SP3 Hell.
The specific files to be deleted or overwritten are application configuration files with an .ini extension. A hold-over from MS/DOS & Win95 days, these files are used in the proprietary application to configure the ODBC handle used for COM data connectors. They are typically set as read-only. What makes it aggravating is that even if the files are set as read-write, they cannot be deleted or overwritten using the VS2005 VB.Net program even when the files are setup initially as NOT read-only.
The applications are installed using an application that maps a before and after image of the system and creates a difference file executable that “installs” the corporate applications.
I've already determined that if I change the files ACL by using VBScript (5.6) to read-write, the VS2005 VB.Net program will work without throwing an exception.
If I change the read-only permission to read-write using Windows Explorer properties applet, the VB.Net program works.
Here's the one that takes the cake, if I run the VB.Net program on an image created with SP2 slipstreamed onto the same volume license image, the VB.Net program works! Even with the files set as read-only. No exception errors.
So what changed in SP3 to break VS2005 VB.Net?
The user logon at both times (during the previous “rename-the-computer” boot and the “current” boot-up) are the same logged-on user who's an Administrator group member. Just the local “domain” name, the computer name has changed. The users have remained the same, the file permissions and the group permissions have remained the same etc.
In all cases, the computer has remained in Workgroup mode. There has been no attempt, at any time, to join a NT or Active Directory domain performed on the computer OS image load.
DEP is on only for Windows OS files.
The OS image starts out as a ghost image, that was a SP3 slipstreamed volume license of Windows XP Pro SP1. In other words, its running as a Windows Pro SP3 system. It operates just fine, no errors, no BSODs etc. You've already heard about the Router-SP3 Hell.
The specific files to be deleted or overwritten are application configuration files with an .ini extension. A hold-over from MS/DOS & Win95 days, these files are used in the proprietary application to configure the ODBC handle used for COM data connectors. They are typically set as read-only. What makes it aggravating is that even if the files are set as read-write, they cannot be deleted or overwritten using the VS2005 VB.Net program even when the files are setup initially as NOT read-only.
The applications are installed using an application that maps a before and after image of the system and creates a difference file executable that “installs” the corporate applications.
I've already determined that if I change the files ACL by using VBScript (5.6) to read-write, the VS2005 VB.Net program will work without throwing an exception.
If I change the read-only permission to read-write using Windows Explorer properties applet, the VB.Net program works.
Here's the one that takes the cake, if I run the VB.Net program on an image created with SP2 slipstreamed onto the same volume license image, the VB.Net program works! Even with the files set as read-only. No exception errors.
So what changed in SP3 to break VS2005 VB.Net?
Saturday 11 October 2008, 3:41 AM
Tempest Toast, Apples and Atoms
A lot of people got clobbered by IKE more severely than my family and I so I should not complain but it has been very busy since then. My project at work got shoved to the back of my "attention stack" once IKE hit. We had 11 days without standard electrical service. You don't appreciate an electric clothes dryer until you have to hang clothes on a line for a few days! Any way, the work I get paid for has now been caught up, instead of the tree clearing and fence repairs I don't get paid for!
My seemingly never ending project finally gets delivered to the internal customer Monday. Besides discovering how crappy commercial software from a couple of International vendors is, it has been a enlightening experience discovering more weirdnesses in Windows XP Pro I didn't already know about.
It is a mix of VB.Net software and a couple of commercial applications that I really wish I didn't have to use. The project delivers a bare-metal hard drive and system restoration that can be operated by screwdriver jockies with absolutely NO computer training besides how to click a mouse. The operators need know just 5 pieces of information!
The next 2 projects are going to be "fun" projects, an embedded Linux operating system image for a touchscreen computer and a Windows XP Embedded image for the same system. So I get to do things at work I've only been able to do at home before this! I have an advantage in that the WinXP Embedded image will be just a modification of a system image currently in use. Of course I'll use Debian 4.0 (etch) for the Linux image.
In any case, we'll get to do an Apples & Apples comparison. Both systems will be running the same Java application running as a user interface and another Java application that operates as a data server to the display app. Same CPU, RAM, compact flash drive, speed, etc. I'm just doing the OS images, not the Java. I'm a bit scorched on programming right now.
Intel has apparently straightened out whatever bugs they had on the low power Atom. I've had a couple of request for quotes come in from vendors that swear that they have the Industrial temperature range part AND the North/South-bridge SCH chip on the same CPU board.
We build a certain amount of our own systems out of single board computers but I've run into something that I don't really understand happening in the Asian manufacturing plants. They will take an industrial temperature (-40 to +85C) rated CPU part (Intel Z530 Atom) and then "marry" it to a bunch of parts that aren't! The resulting board can't withstand the high temperatures and they think that's OK?
I understand using an Industrial temp part if that's the only one offered by Intel (as the Atom Z530 is) but then not to offer at the very least a tested version of the SBC for an industrial temp range? What marketing geniuses came up with that as a sales strategy?
The Z530 part was specifically designed for the Industrial and Automotive market yet very few of the manufacturers are offering a board that takes full advantage of the part.
I actually had a vendor promise me a quote on an Industrial temp range single board computer based on the Atom and then sent me a spec sheet that indicated the CPU would quit working at 60C. When I brought that to their attention, they said they decided to change the specifications because no one needed the expanded temperature range!
I typically "bake" our computers at 122F (+50C) and see how long they will run before they crap out. The systems that can run for a couple of weeks, we'll buy. Some of our customers drill where there is a lot of sand and a lot of sun to keep it hot. So yes heat is an issue and we DO buy the extended temperature range systems!
My seemingly never ending project finally gets delivered to the internal customer Monday. Besides discovering how crappy commercial software from a couple of International vendors is, it has been a enlightening experience discovering more weirdnesses in Windows XP Pro I didn't already know about.
It is a mix of VB.Net software and a couple of commercial applications that I really wish I didn't have to use. The project delivers a bare-metal hard drive and system restoration that can be operated by screwdriver jockies with absolutely NO computer training besides how to click a mouse. The operators need know just 5 pieces of information!
The next 2 projects are going to be "fun" projects, an embedded Linux operating system image for a touchscreen computer and a Windows XP Embedded image for the same system. So I get to do things at work I've only been able to do at home before this! I have an advantage in that the WinXP Embedded image will be just a modification of a system image currently in use. Of course I'll use Debian 4.0 (etch) for the Linux image.
In any case, we'll get to do an Apples & Apples comparison. Both systems will be running the same Java application running as a user interface and another Java application that operates as a data server to the display app. Same CPU, RAM, compact flash drive, speed, etc. I'm just doing the OS images, not the Java. I'm a bit scorched on programming right now.
Intel has apparently straightened out whatever bugs they had on the low power Atom. I've had a couple of request for quotes come in from vendors that swear that they have the Industrial temperature range part AND the North/South-bridge SCH chip on the same CPU board.
We build a certain amount of our own systems out of single board computers but I've run into something that I don't really understand happening in the Asian manufacturing plants. They will take an industrial temperature (-40 to +85C) rated CPU part (Intel Z530 Atom) and then "marry" it to a bunch of parts that aren't! The resulting board can't withstand the high temperatures and they think that's OK?
I understand using an Industrial temp part if that's the only one offered by Intel (as the Atom Z530 is) but then not to offer at the very least a tested version of the SBC for an industrial temp range? What marketing geniuses came up with that as a sales strategy?
The Z530 part was specifically designed for the Industrial and Automotive market yet very few of the manufacturers are offering a board that takes full advantage of the part.
I actually had a vendor promise me a quote on an Industrial temp range single board computer based on the Atom and then sent me a spec sheet that indicated the CPU would quit working at 60C. When I brought that to their attention, they said they decided to change the specifications because no one needed the expanded temperature range!
I typically "bake" our computers at 122F (+50C) and see how long they will run before they crap out. The systems that can run for a couple of weeks, we'll buy. Some of our customers drill where there is a lot of sand and a lot of sun to keep it hot. So yes heat is an issue and we DO buy the extended temperature range systems!


