* Posts by Trog

4 publicly visible posts • joined 19 Apr 2007

WiMAX has 'failed miserably'

Trog

Latency

Lance wrote "Oh, 3G also has more latency than EDGE/GPRS"

Not true. GPRS is about 700ms round-trip ping times, Rel99 3G about 200ms, and HSDPA about 130ms. Try it by pinging a server if you don't believe me.

There is a difference in their cold-start peformance - 3G escalates the connection speed through a number of steps each time upgrading the mode and bandwidth. These transitions can take a 1 or 2 seconds to complete - and so for a small amount of data such as a very simple webpage GPRS can beat 3G - but it quickly loses this advantage once the download exceeds a few tens of kilobytes.

Very surprised to learn Buzz's WiMax had a latency of ~1second. Sounds like a problem specific to their network (congestion, overloaded equipment?) unless others can confirm this is generally true for Wimax?

Cambridge brain touts wind-n-server combo farms

Trog
Unhappy

Unconvinced

I'm unconvinced this is actually more environmentally friendly.

Waste heat in a server-farm usually needs more energy use in terms of air-conditioning to get rid off. Waste heat in people's homes make a small contribution to their central heating and so arguably is not waste heat ((in winter at least).

Power is transported at very high voltages to reduce transmission losses (which are most dependent on current flow). Is it really more power efficient to move many more low-voltage bits around, including all the processing and electricity usage of all the routers and switches en route?

Trog

Study: Users pay for Microsoft patent woes

Trog

RE: lawsuits Not royalties... and other things

Simon,

'this calculation is based on "patent-related lawsuits and settlements." not royalty payments, which are a valid part of the OS if included. If M$ want to pass the charge on to customers, then they can, and nobody is arguing with that.'

I think people are arguing with that. If the charges they've added related purely to legal fees and fines ('punitive damages'), then I'd agree with you. However I don't think the charges do - for instance the report cites M$ settled with InterTrust which was announced as a single one-time payment of $440m granting a perpertual license for InterTrust's patents in Microsoft's technology.

Unless the lawsuits is settled publicly in court there's often no breakdown of how much of an out-of-court settlement is punitive and how much is for royalties past and future and so there's simply no way of splitting the money between necessary royalty payments and unnecessary punitive and legal costs.

Trog

Re: Yes you are right

"Yes of course that is how business works. The point is that it is the customer who pays for Microsoft breaking the law. That is the point.

Such recuperation is not part of any other OS and therefore this is a negative for Windows."

Microsoft aren't been charged patent royalties because they are breaking the law, it's because the patent owner invented something they want or need to make use of.

For instance, Windows Media needs to license patents for their codecs and their DRM. If the Linux community developed a similar DRM solution or included codecs such as H.264 or MP3, they would also have to pay for the patents too. Developing something as open-source or in a 'clean room' environment doesn't make it exempt from patent law (it only really helps with copyright law).