Posted 11/01/2009 at 08:24:09am

I thought the same thing when I first read 100kg, but you have to consider 2 things:

1) By 2015, the machine will probably be capable of lifting a good bit more.  If not, improvements (and competition) should begin to appear shortly after its commercial release.

2) You may be able to lift 100kg to a full press (8' high?) if you're a relatively lean/muscular 200+, since indications are that the world record is over 500 pounds (this is with some snap, mind you, not a clean military-style press).  However, I doubt you can do it with much more than a barbell, and I doubt you can do it for as long as anyone using the machine could.

You can lift for a while, but you get tired, especially if you lift near your maximum weight.  When you get tired, you rest -- but you still get paid.  If someone in the machine can lift the machine's maximum with little or no rest like a crane driver might, his efficiency will pay for the machine over time.  Also, let's not forget that this machine will likely reduce lifting injuries on the job, which means less lost time and less spent on worker's compensation and disability.

By 2020, when the price has come down a bit, I bet every major construction site has as many of these as they do Bobcats.

Posted 11/01/2009 at 07:48:21am

DX10 wasn't the improvement everyone hoped for (much like Vista), but it's also fair to say developers didn't spend enough time with it to use it as efficiently as possible.  DX9 is well-known to developers, who have had plenty of time to learn to use it for realistic graphics that run relatively fast on even somewhat older DX9 hardware. 

The idea of developers taking years to gain efficiency can be seen in multi-core (parallel) processing -- users should see near 1:1 speedups for every core added in computation-heavy apps, but in reality users are rarely seeing speedups of more than 50%.  Multi-core CPUs are just making multi-tasking faster so far.  Parallel processing has been around for a couple of decades, but commercial developers have only had an incentive to move away from sequential processing in the last few years.

Likewise, it should take a year or two before we really see what DirectX 11 is capable of in real world applications (test/demo applications are easier).  That said, DX11 does come with some nice features: 

1) It takes advantage of the unified shader architecture introduced in DX10 (which even sped up DX9 games).  It should, arguably, take better advantage of the architecture than DX10 did due to improvements which should have been made in the low-level libraries.  By all indications, this should take little or no effort on the part of the game/application developers -- which is a good thing for us, as not every developer is as adept at understanding hardware architecture's effect on their code.

2) It adds support for "new" hardware tessellation (AMD technically already had tessellation units, but built-in support will make this more useful to everyone).  Tessellation, the process of breaking down complex shapes into smaller, simpler shapes (triangles, in this instance), is a relatively intensive process that can be made much more efficient and effective by using dedicated hardware designed to handle it.  For the developers, it means less work and better results.  For us, it means more detailed graphics that take less time to generate (less time per frame means more frames per second).

3) It brings better support for compute shaders.  While the concept has existed for a while, DX11 brings with it explicit compute shader support.  For one, this means that GPUs can be used more efficiently for general purpose applications (think video transcoding).  For another, it means that both developers and hardware manufacturers have more freedom in how to apply algorithms to objects and data sets.  This means fancy details like motion blur, soft shadows, depth of field, and other post-processing graphical enhancements will again look better and run faster.  Also, since compute shaders are more mutli-purpose, you could see improvements in AI or physics processing if your card has enough processing power to handle the extra duties.

So, in short, it will take time for developers to use DX11 to its best, and probably a couple of patches for Microsoft to weed out the remaining bugs.  But there are improvements over DX10 in DX11, including some that software developers -- both the game-making type and more real-world type -- may get excited about this time.

Posted 11/01/2009 at 02:59:09am

I'm really not trying to be a jerk on purpose here, but on the planet I live on Windows 7 DID get a late 2009 release.

Also, being a software developer and knowing just a little about Microsoft's development cycles, neither Vista DX11 support nor Windows Mobile development are going to affect Windows 7 development.  Different teams work on different products.  If a product is dropped, a few of those people will fill empty slots in other teams, but some of them will also just go work for a different company.  That is, Microsoft isn't dividing it's manpower -- it hires more people to develop more products.  That's how business works.

Posted 10/25/2009 at 11:14:13pm

I would just like to point out that 110V very much can kill you -- it's just unlikely to happen.  Anything under about 1000V will cause your muscles to contract, which is why it feels like getting punched in the chest (or hit by a baseball bat if it's 220V or more).  1000V+ causes the opposite reaction, which is why you get "knocked away".

It's this muscle contraction that can result in your death.  If you grabbed a live wire/rail/etc at 110V, your hand would clamp down, and you could quickly top out a 15-20A breaker.  The breaker *should* trip before you would even be rendered unconscious, but is not designed specifically to trip due to electrocution (GFCI is more reliable for that, but still not fool-proof).  According to US Consumer Product Safety and the National Institute for Occupational Safety and Health, about 411 people died from electrocution in 2001.  19% of those were due to large appliances and 11% were due to "installed household wiring", so we can estimate that about 100 people were killed by 110V current (if we assume some of those were 220V appliances).  That's still less than 1 in a million per year, but it does happen.

However, since we tend to touch things with the tips of our fingers (especially if you know what you're doing), usually you just get a good half-second jolt, cuss a bit, and move on.  Skilled electricians know and inherently trust simple facts like if you create a closed circuit, you can twist the hot wires together safely by hand.  However, since most people are *not* skilled electricians, they should *not* throw caution to the wind -- never work with hot (electrified) wires unless you know exactly what you're doing.  Even skilled electricians are supposed to "unground" or insulate themselves when working hot according to OSHA guidelines.

I don't mean to argue or be a stick in the mud.  I just don't want people to get the idea that 110V is safe and that they don't have to take precautions.  Always practice electrical safety, folks! Plus, getting popped still hurts like a SOB even if it doesn't injure you.

Posted 10/24/2009 at 03:13:15am

I have to support Micro-USB over Mini-USB.  Mini-USB just isn't "mini" enough to save space on the smallest devices.  I think it should be mandatory (I'm pretty sure South Korea did this already), and I support it for more than just cell phones.  I think any device designed to be a mobile consumer device should use Micro-USB, so you can find a charger anywhere and charge all your devices using the same adapter.

For those that already have lots of Mini-USB cables and AC/DC adapters, you can pick up a small Mini-to-Micro USB adapter for under $2.

Posted 10/22/2009 at 01:18:57am

More specifically, it comes from the naming convention of the Intel 80x86 line of processors.  When you ran the old 286, 386, and 486 PCs, they were based on the 80286, 80386, and 80486 architectures, which were backwards compatible with the Intel 8086 architecture's instruction set.  When the Intel 80586 architecture was introduced, Intel chose to label it "Pentium" for trademarking purposes.

Also, when people say "x86", they are referring to x86-32. The original Intel 8086 was 16 bit, which would be x86-16.  However, calling it x86-16 is really back-labelling -- the Intel 80386 was a 32-bit architecture, and is where the x86 moniker began.

And now you know!

Posted 10/22/2009 at 12:40:39am

Do you read the magazine and/or website?  Norton 2009 was considered one of the best (if not THE best) boxed Antivirus apps this year.

Granted, there are multiple antivirus and Internet security apps in that list, though, which is pretty useless.

Posted 09/30/2009 at 03:58:21am

See the following review: Antec 900 Two
Just bought one for a Core i7-860 based machine I'm building.  The top fan is bigger than my hand at full spread.  Can't comment on how easy it is to install in yet, as I'm still in the prep stage of the build. 

Posted 09/17/2009 at 06:45:19pm

That would defeat the purpose of patents.  Let's say I come up with this great design for a new vehicle feature (not just an idea -- a design), but prototyping it and perfecting it will cost around $500K.  Then I take it to Chrysler to sell it to them.  Under your system, they can just run with it and build it before I can and the patent (that I funded) belongs to them.  Heck, under your system, the first person or company to get one built and marketed would own the patent.  Since patents must be made public in order to be fair and useful, simply patenting something would be giving it away.

The problem is not the system, it's the patent office.  They keep forgetting that patents are there to protect the public domain and bring out innovation -- that's why patents expire.  Furthermore, they keep granting patents for ideas, which are not patentable.  Being the first to patent the idea to connect 10 computers together -- not valid.  Being the first to invent a network card capable of sharing information and the design and equipment for a ring topology -- valid patent, but only if YOU did it and weren't just the first person to figure out it wasn't patented (making it public domain).

Posted 09/13/2009 at 05:23:37pm

I'm willing to bet this was an article either in the magazine or meant for the magazine (haven't seen that article yet if it's it this month's).  That format is actually quite print-friendly.  It just doesn't look as good on a computer screen.

As much as MaximumPC wants to be more online than magazine, they are still first and foremost a magazine right now.  I love having the website, including all the major reviews and coverage from the magazine, right at my fingertips.  They are still perfecting their online side.  I'm willing to sacrifice web-friendly charts if they keep up the high quality of the magazine.  I still haven't found a magazine I like quite as much, mostly due to lack of focus and letting magazines become advertisement books with the occasional article in them.