Channel 9 - Discussions by msemack

Coffeehouse - Do strimmers hold more promise than they deliver?

msemack — Fri, 11 Jun 2010 16:45:50 GMT

Ian2 said:

msemack said:

*snip*

This looks like the closest match over here:

http://www.blackanddecker.co.uk/gardentools/productdetails/catno/GLC2500/

I've had a few petrol based so maybe its time to try something new ..

I have seen some newer cordless trimmer that use Lithium Batteries or are higher voltage. They might be more compelling:

I have also seen similar Lithium Battery models from the other major power tool companies at our local Home Depot.

For me personally, I use some other tools that share the same battery (http://www.blackanddecker.com/ProductGuide/CategoryOverview.aspx?cPath=1498.4945). So I probably wouldn't upgrade my trimmer unless I was upgrading everything. Maybe once B&D comes out with Lithium equivalents of all of them.

As I said before, the battery power definitely has it's limits. When the battery gets low, it will start to bog and slow down. I only use it for a total of <30min after mowing to tidy up. If I had a bigger yard or fancy landscaping to maintain, I would probably use a gas trimmer.

Coffeehouse - Do strimmers hold more promise than they deliver?

msemack — Fri, 11 Jun 2010 13:30:12 GMT

Never heard the term "strimmer" before. I've heard it called "weed whacker" and "string trimmer".

I pull mine out probably every other time I mow. I use it to trim the grass right at the edge of the house foundation, in that <1inch gap the mower won't touch. We also have a kid's play set in the back yard that I have to trim around. (Eventually I want to put down mulch or some kind of edging around the house so I don't need to do this.)

Probably the biggest use for my trimmer is for along the edge of the sidewalk and driveway. My trimmer converts into an edger by spinning the head around so it trims vertically. It does a great job giving the sidewalk and driveway a "fresh from the barber" look.

Mine feeds the string automatically when it gets to short. Never had to even think about it. Crud, now I wonder how much string is left in it.

This is the model of trimmer I have, BTW. http://www.blackanddecker.com/ProductGuide/Product-Details.aspx?ProductID=23463

The battery power definitely has its limits, but it is capable enough for my lot (0.35 acres). I have a few other Black and Decker yard tools that use the same batteries, so it is easy to interchange, and I always have a spare fully-charged battery ready to go. The battery is nice because I don't need to drag a cord around, and it doesn't need the care/maintenance that a gas engine needs.

Coffeehouse - Secure IM

msemack — Fri, 08 Jan 2010 14:04:42 GMT

Isn't Skype encrypted?

Coffeehouse - Skype wants your BIOS

msemack — Wed, 07 Feb 2007 20:54:54 GMT

AndyC wrote:

Er, they are supposed to be. Networks could seriously break if they weren't.

Welcome to the real world. MAC addresses usually are unique, but not always.

1. Most OSes let you spoof MAC addresses.

2. Most ethernet controller chips let you modify the MAC address at runtime (a requirement for #1).

3. Many broadband routers allow MAC address spoofing.

4. Some ethernet controller chips have a bug where some of the MAC address bits will periodically change. This bug was present on an extremely popular PCI Ethernet card sold at retail chains across the country.

5. Some el-cheapo network card vendors will recycle MAC addresses, rather than purchasing a new block of numbers.

6. Early Sun Microsystems workstations had dual ethernet ports, but allocated MAC addresses PER WORKSTATION, not per port. So, there were 2 ethernet ports with the same MAC address.

Coffeehouse - Skype wants your BIOS

msemack — Wed, 07 Feb 2007 20:35:35 GMT

ScanIAm wrote:

Why can't they be used as a universal identifier?

Not sure if this answers the question, but MAC addresses are not universally unique.

Tech Off - XP service pack 2 and mup.sys

msemack — Fri, 16 Jun 2006 02:46:50 GMT

Even if you reinstall, you still haven't isolated the cause of the problem.

The 0x7B stop code and the corrupt registry file (\system32\config\system) point to a failing device in your storage subsystem. Check your hard disk and IDE cables.

Your hard drive maker probably offers a drive test utility as a free download. Get a copy and run it. Make sure there are no bad blocks on the drive.

Also, double check your IDE cables. Make sure they are all 80-conductor. Also check cable lengths. Any IDE cable longer than 18" is bad and should be thrown out immediately. Likewise for those sliced/narrow IDE cables.

It may also be your IDE controller, especially if you are using a cheapo IDE RAID controller.

Another oustide possibility is that it's bad memory. It may be worth running an overnight cycle of Memtest86.

Coffeehouse - Windows Explorer security question

msemack — Mon, 08 May 2006 13:53:35 GMT

rjdohnert wrote:

If I am running an alternative shell and not Windows Explorer and Im using a different Filemanger than Windows Explorer, would I still be as vulnerable as users who are running it as their default shell?

It depends. Some alternative shells hook into the same libraries of Windows as Explorer. If a vulnerability is found in one of those libraries, you're just as vulnerable.

You'd have to look at them on a case-by-case basis.

Use depends.exe to look at DLL dependencies for Explorer (look at explorer.exe). Then, compare it to your alternate shell of choice. If you see any common dependencies, then a vulnerability in one of those DLLs could affect both.

Tech Off - Process 'SYSTEM' Run By 'SYSTEM' Using 100% CPU

msemack — Thu, 04 May 2006 00:47:12 GMT

anthq11 wrote:

Is there any causes of System (not System Idle Process) running at 90-100 for extended periods of time?

Most likely it's a buggy driver. USB devices are particularly notorious for having drivers with a high CPU load.

I suggest that you shut off or remove devices one at a time until you find the culprit. Start with USB devices.

Coffeehouse - Richard Stallman just spoke to me

msemack — Fri, 28 Apr 2006 16:35:47 GMT

IRenderable wrote:

He is a damn good programmer.

What exactly do you base that on? I'm not trolling, it's a genuine question.

Have you reviewed code that he has written?

Have you discussed software architecture with him?

Coffeehouse - Starting Windows over...

msemack — Sat, 22 Apr 2006 02:43:53 GMT

messerschmitt wrote:

Read my previous post.

Your previous post does not answer my question.

How will ditching backward compatibility make my life better?

All it will do is require me to fork out a large sum of money to upgrade all of my applications.

Coffeehouse - Starting Windows over...

msemack — Fri, 21 Apr 2006 21:04:05 GMT

messerschmitt wrote:

I would prefer tough that Microsoft ditch the backwards comp in Vista, including dumping win32, why not go .net all the way?

Why though? What does it gain you?

Furthermore, what does it gain the CUSTOMERS?

Ditching the Win32 stuff DEGRADES the user experience, because people can't run the programs that they know and love. Remember, it's all about the applications.

Coffeehouse - Starting Windows over...

msemack — Fri, 21 Apr 2006 16:54:48 GMT

Most often, re-writing from scratch is a mistake. It seems very tempting, but usually ends up ding more harm than good.

http://www.joelonsoftware.com/articles/fog0000000069.html

You also have to ask yourself: why would you want to develop a new user mode API? What does it actually gain you? And more importantly, what does it actually gain your CUSTOMERS?

Coffeehouse - linux failure to support proprietary drivers

msemack — Tue, 18 Apr 2006 20:10:33 GMT

Erisan wrote:

Hmmh. There has been a lot of discussion about stable driver API in Linux.

And it has been actively opposed by the Linux kernel community, particularly Linus himself.

Basically, I've heard the following arguments from the Linux kernel contributors:

1. Backwards compatibility makes the kernel "impure". If I want to fix a bug in a kernel function, I want to be able to just fix it without having to deal with the "yucky" concerns about backwards compatiblity.

2. If your driver is in the kernel source tree, the API changes are a non-issue because "the community" will keep your driver up-to-date with the latest kernels.

3. Since it's all free software, API changes have no cost.

4. If your driver isn't in the kernel tree, then it must be an evil proprietary driver. If you are a proprietary driver, then I deliberately want to make your life as hard as possible.

Personally, I've found these arguments to be largely philosophical, not very realistic, and almost childish at times (ha ha, I broke your driver). Some of the arguments (like #2 and #3) are simply wrong.

Coffeehouse - linux failure to support proprietary drivers

msemack — Tue, 18 Apr 2006 16:04:47 GMT

imekon,

Under Windows, when a new driver model comes out, old drivers are not immediately broken.

Case and point, I can still run older NT4.0-based drivers (pre-WDM) on Windows XP. Even though Windows 2000 and XP have added newer driver interfaces (like power management and PnP), the old drivers continue to run. The older driver interfaces are not broken.

Even with the new WDF model of Vista, older WDM drivers can still work.

This is something that the Linux kernel community does not do (and refuses to do). They regularly break the existing driver interfaces.

Coffeehouse - linux failure to support proprietary drivers

msemack — Tue, 18 Apr 2006 13:32:06 GMT

The Linux kernel APIs change with almost every patchlevel release of the kernel (e.g. 2.6.4 to 2.6.5). This means that when a new kernel release comes out, all of your Linux kernel modules stop compiling against it.

Point releases of the kernel (e.g. 2.4.x to 2.6.x) bring even more dramatic changes.

To make matters worse, most of the major Linux distributions are using some kind of modified kernel, which may have its own API changes.

To deal with this giant matrix, you have to add lots of #ifdefs to the code (if this kernel version, use this API, unless it's a Redhat kernel, then use this API, etc). It gets pretty complicated pretty quickly.

Because the APIs are a constantly moving target, you need to have developers monitoring kernel releases and constantly doing build tests.

We develop drivers for both Windows and Linux. The Linux driver costs 20-30% more for initial development. On top of that, the recurring maintainence costs for the Linux driver are MUCH higher than the Windows one.

These are costs that we do not have to pay with the Windows design. Even when Microsoft transitions to a new driver model (e.g. WDF/Vista), the older drivers continue to work.

Coffeehouse - Why not AMD mac machine?

msemack — Tue, 11 Apr 2006 17:30:43 GMT

Keep in mind that Apple could easily switch to AMD in the future.

Now that they're making the transition to x86, they have multiple CPU vendors to choose from. All of them are binary-compatible. They can use processors from Intel, AMD, Via, even STMicro.

They could even mix and match (e.g. Use a low-power Via chip in the Mac Mini, and use an AMD Opteron in the their high-end line).

Also, now that they've transitioned, they have more negotiating power. They can play Intel, AMD, and Via off of each other to negotiate for better pricing.

Basic rule in business: Commoditize your complements.

It's why the PC industry is so successful. It's why IBM endorses Linux. It's why Apple offers a not-very-profitable online music store.

http://www.joelonsoftware.com/articles/StrategyLetterV.html

Coffeehouse - OK, so why doesnt vista use EFI to boot?

msemack — Tue, 11 Apr 2006 13:23:17 GMT

MalEbenSo wrote:

Read that again: The same driver for all versions of Windows, Linux, OS X, BSD! Can you imagine what a boost that would mean to roll-out new hardware and stabilize old hardware?

The Wikipedia article is misleading on it's concept of OS-neutral drivers. Don't put a lot of faith into it.

Coffeehouse - OK, so why doesnt vista use EFI to boot?

msemack — Tue, 11 Apr 2006 13:16:21 GMT

Despite being a current industry buzzword, the advantages of EFI are practically non-existant to the end user. EFI have been around for years. Motherboard adoption has been near zero. There's a reason for that.

Most people (even on this forum) wouldn't be able to tell the difference between an EFI motherboard and a non-EFI one. The setup menus look the same. Your system will run the same (no faster or slower).

The preboot shell is only useful is a few (very rare) situations. It is really targeted at embedded systems vendors who still use DOS. It's not very useful for the consumer.

Certainly it would be nice if Microsoft added support for EFI to the 32-bit Windows builds (without axing BIOS support). You can't go wrong with more features. However, this is hardly a show stopper. It makes more sense to dedicate manpower to more useful features.

Furthermore, the lack of existing EFI motherboards means that any EFI support in Windows will have minimal real-world testing. This makes it very likely there will be bugs in the support, which will have to be fixed later with a service pack.

You have to understand, BIOS doesn't really "hold back" the PC architecture. Most OSes totally ignore the firmware once they boot (BIOS or EFI). EFI doesn't really make anything possible that isn't already possible.

For those who are saying that EFI is connected to DRM: You are way off the mark. There is nothing special about EFI to enable DRM support. Certainly you can use a TPM chip on an EFI motherboard. However, you can use a TPM chip on a BIOS-based motherboard. My Thinkpads have had TPM chips for years.

Coffeehouse - Backwards Compatability...Is There a Point Anymore?

msemack — Thu, 23 Mar 2006 20:31:10 GMT

All of those legacy applications represent BILLIONS (if not trillions) of spent dollars in R&D. Telling a company to disregard that investment because you don't feel like supporting it anymore is a sure way to alienate your customers.

It's not a coincidence that the most successful computing platform (the wintel PC) is the only one that can run software that was written in the 80's.

As it stands now, there is an uproar whenever Microsoft breaks application comptibility. And they try harder than any other SW company to maintain backwards compatiblity.

You have to understand that there is HUGE amount of software out there that CAN'T be upgraded for a variety of practical/economic reasons. Here are some real-world examples:

1. An in-house application that was hacked together, possibly by a non-programmer who "knows about computers". That program is now years old. The guy has left the company. The company may not even have the source code for it anymore.

2. The company that wrote the application is no longer around.

3. You may have the source code, but you can't locate a copy of the now-discontinued compiler.

4. You may have the source code, but it's written in COBOL and no one is around that knows COBOL anymore.

5. There may be a new version of the application available, but it uses an incompatible data format. You can't upgrade without converting your data over (which isn't always possible).

The application in the above examples may be something critical to the company. Immagine you are a billion-dollar company with a custom in-house accounting program. You want to upgrade from Windows 2000 to XP, but the accounting program won't work under XP.

To fix this, you have two options:

1. Fix the application (which may be impossible for the reasons described above).

2. Not upgrade to XP. This means fewer sales for XP, and less money for Microsoft.

You see, it is in Microsoft's best interest financially to make the code as backwards-compatible as possible.

It is also in YOUR best interest as a USER for Microsoft to do this. It means that you have more freedom to upgrade your copy of Windows without throwing out all of your old software.

Tech Off - "Windows - Registry Recovery"

msemack — Sat, 18 Mar 2006 15:16:21 GMT

My biggest suspects:

- Bad memory

- Dying hard drive

Coffeehouse - Flash Drive Laptop Exist

msemack — Sun, 12 Mar 2006 04:40:25 GMT

Shining Arcanine wrote:

You are looking at the chips USB Keys use and are making a bunch of generalizations.

No I'm not. As I said earlier, a large number of the customers I deal with purchase embedded systems that USE 2.5" flash drives. We actually pre-install 2.5" flash drives from M-Systems, Memtech, and others.

Shining Arcanine wrote:

1. Capacity is not a major problem when a 2.5" flash drive can contain dozens of chips.

2.5" drives have a practical height restriction. This means you can only fit a single PCB inside the metal casing. You'd have a very hard time cramming 10 or more chips (assuming standard 48-pin TSOP flash pacakges), into a height-compliant (9.5mm) drive casing. I have never seen a height-compliant flash drive containing more than 8 flash chips.

The selection of flash drives drops off considerably once you get above 4GB. There are companies that offer 2.5" drives up to 128GB, but be prepared to pay a hefty premium (several thousand dollars). They also tend to be much taller than the standard laptop drive.

Shining Arcanine wrote:

2. Prices for flash memory are going down. I treated myself a flash drive late last year and now I can buy the same model for half their price.

I never said that prices weren't going down (just about any piece of computer hardware gets cheaper by the day). However, traditional rotating disks are dropping in price also. For a flash-based drive to become cost-effective, it has to drop in price FASTER than its rotating equivalent. At the current rates, the per-megabyte price gap is WIDENING. This means flash drives will always cost a premium, even several years from now. In fact, the premium will get WORSE, not better over time.

Shining Arcanine wrote:

3. When you RAID 0, say a dozen chips, throughput is not a problem and modern single chip USB Keys can get 20MB/s or over, so if the drives were only to have a few chips, throughput would not be a huge problem.

Sounds nice in theory, but you need a huge amount of parallelism to make up for the limitations of flash devices. Rotating 2.5" drives consistently come out higher on sustained transfer rate benchmarks than their flash cousins.

Shining Arcanine wrote:

4. While each bit of a flash drive can only last a 100,000 writes, with special algorithms modern flash drives move static data to heavily used areas and often changed data to lightly used areas so the write limit in practice is actually billions of writes; at least according to Samsung anyway.

Did you not read my post? I already mentioned wear-leveling. I siad it EXTENDS the life of flash drives, but DOES NOT eliminate the problem. A typical Windows install on a flash drive (used every day) can wear it out within two years.

Shining Arcanine wrote:

At this rate we will have 64GB for $900 in 2007, 128GB for $900 in 2008, 256GB for $900 in 2009, 512GB for $900 in 2010 and so on. That should halve the prices for the 32GB model each consecutive year, meaning $450 in 2007, $225 in 2008, $112.5 in 2009 and $56.25 in 2010.

Wow, I can pay almost $1,000 for a drive that is considerably smaller than current rotating drives. If I wait until 2010 it will be cheap. However, by 2010, it will be pitifully small (a next-gen DVD drive will have more capacity). By 2010, laptops will probably be shipping with >500GB drives.

Don't assume that your storage requirements will remain constant for the next four years. Newer versions of software (Vista, Office 2007) will have higher disk space requirements. Also, the amount of data that you need to store will probably grow. Your music/movie collection will get probably get larger. Games get more detailed graphics and textures.

Coffeehouse - Flash Drive Laptop Exist

msemack — Sat, 11 Mar 2006 23:19:30 GMT

Shining Arcanine wrote:

Hopefully in a few years the price of 32GB flash 2.5" hard drives will go down to affordable levels.

Don't hold your breath. I follow the flash drive market pretty closely.

High-capacity flash drives have always been pricey. And the per-megabyte price gap for flash vs. rotating drives has been getting WIDER, not narrower.

Hard drive per-megabyte prices have been dropping a lot faster than flash drives.

Coffeehouse - Flash Drive Laptop Exist

msemack — Sat, 11 Mar 2006 23:13:52 GMT

2.5" IDE flash drives are pretty common. A lot of the ruggedized systems we sell ship with them (extended temperature versions, of course).

http://froogle.google.com/froogle?q=2.5-inch+IDE+flash+drive&btnG=Search

Alternatively, you can also use a CompactFlash card as your laptop's hard drive with an inexpensive adapter.

http://siliconkit.dnsalias.com/cart/tcf44.html

However, flash drives have a few drawbacks:

1. Capactity. Flash drives are usually in the <1GB range.

2. Price. They cost a lot more than an equivalent rotating drive, especially in >1GB capacities. For a 4GB flash drive, expect to pay almost as much as the laptop.

3. Performance. Flash drives have zero seek latency, but their throughput is terrible.

4. Lifespan. Flash memory can only survive a finite number of erasure cycles. The exact number varies depending on the specific flash chip, but it is FAR below the expected MTBF for a hard drive. A lot of flash drive have wear leveling technology to extend their lifespan, but this does not eliminate the problem.

For general consumer use, something like this seems far more appealing:
http://digital-lifestyles.info/display_page.asp?section=platforms&id=2146

Tech Off - Regarding Windows Registry performance

msemack — Mon, 06 Mar 2006 16:43:40 GMT

Prove that the slowdowns you are seeing are caused by "registry fragmentation".

Coffeehouse - Microsoft needs to quit stalling

msemack — Wed, 01 Mar 2006 21:07:31 GMT

Do you want it done FAST, or do you want it done PROPERLY? Good software takes time, even for a company with Microsoft's resources.
It's not even a question of hiring more developers, or assigning more people to a project. You can't make a baby in one month by putting 9 women on the job.