This is Peter Amstutz's Typepad Profile.
Join Typepad and start following Peter Amstutz's activity
Join Now!
Already a member? Sign In
Peter Amstutz
Recent Activity
I think it's perfectly legitimate for a regulatory body enforce good engineering practice; you don't want to drive over a bridge or live in a house that isn't up to code. Computer networks are increasingly considered critical infrastructure, why shouldn't it be regulated where the public interest is concerned? [*] [*] Of course there is a slippery slope argument here, so let me be clear I am 100% against any kind of censorship by government or corporation. An interesting special case that hasn't really been mentioned is network security. Verizon DSL blocks outgoing port 25 to fight spam botnets; however this is incredibly annoying if you want to use a mail agent via a legitimate relay. Personally I think they should not be allowed to block ports without cause, but should be able to disconnect customers who have been pwned.
Toggle Commented Feb 15, 2011 on The Importance of Net Neutrality at Coding Horror
Net neutrality, like climate change, is one of those political topics that is a documented target of astroturf campaigns by incumbent players (Comcast, AT&T) who want to ensure the gravy train keeps rolling: So take any hysterical anti-regulation comments with a grain of salt. Net Neutrality is about preventing ISPs from engaging in various monopolistic practices. This is about consumer protection. Nobody is suggesting that they be prevented from managing their networks, but ISPs should provide all bits that customers request with best effort and not deliberately slow-walk competitors services. People seem to forget that the marginal cost of providing bits is very near to zero; once the infrastructure is built and the network admins get paid, it doesn't cost anything more if your network is running at 5% or 50% utilization. If the network is at 5% utilization and the ISP still decides to throttle a competitor's service, that is a monopolist using their position to strangle innovation. That is what net neutrality is intended to address.
Toggle Commented Feb 15, 2011 on The Importance of Net Neutrality at Coding Horror
I'd also point out that the alleged trade off between "spend more time writing better code" and "just buy more memory" could in some cases be an effort to hide problems like memory leaks, where you are just delaying the inevitable. Disk speed is primarily an issue for latency-sensitive operations (like interactive database queries) but apps that have to churn through a lot of data ought to be batching up reads and issuing async I/O requests so that the OS can fetch or copy the data independently from the application. The need for programs to be mindful of its current working set doesn't go away with more RAM, it just moves up the cache hierarchy to the CPU.
This is an interesting twist on the software monoculture problem. By being the overwhelming favorite, Google gives spammers a single target to focus on. On the other hand, unlike viruses, there are no inherent platform differences that prevent spammers from tweaking their content scrapers to poison other search engines as well. So if Bing were to gain more market share its results would likely start to drown in noise as well unless they have some secret sauce (or armies of content reviewers) to defend the walls against the barbarian hordes of spammers. The problem with adding a button that says "this web site is useful" or "report abuse" is that it shifts the battle to gaming that metric instead. The other problem is that from the algorithm's perspective, information is information, so who cares if some content scraper serves up information copied from someone else as long as you the searcher get the answer you're looking for, right? In some cases, the sites in question could be legitimate mirrors. When discussing flaws in the ranking algorithm, this goes to the heart of how you phrase the question -- assuming there are no flaws in the software, it is probably performing as intended, and this is essentially a garbage in/garbage out problem. Unfortunately, there is a lot of garbage on the Internet.
Toggle Commented Jan 3, 2011 on Trouble In the House of Google at Coding Horror
My understanding is that the evolution of display connectors is something like this: Originally there was VGA, which is a purely analog RGB signal. DVI is essentially a digital version of a VGA signal, where analog voltage changes are replaced with digital values. This is due to the rise in prominence of digital flat panel displays, where the digital signal is "native" to the underlying hardware in the same way that an analog signal is "native" to a CRT. DVI also includes pins for analog VGA signal for backwards compatibility. HDMI is basically DVI with analog VGA compatibility removed, audio channels added, and a more consumer-friendly (some would say flimsy) connector. DVI and HDMI are nearing the end of their evolution because the technology cannot support further increases in the maximum resolution. DisplayPort is significantly different. It is a high bandwidth packetized digital signal with significantly more headroom to support higher maximum resolutions. Also the packet approach allows for alternate transmission media like fiber optics, as well as additional data streams such as multiple video signals, audio streams, or even internet packets. That said, it is going to take awhile for DisplayPort to gather momentum. Considering how long it took for USB to be fully established as the standard connector for input devices (introduced 1996, computers as late as 2002 often still only had a single USB port and came with PS/2 mice), new standards take a long time to displace the existing installed base.
Toggle Commented Apr 8, 2010 on Three Monitors For Every User at Coding Horror
Peter Amstutz is now following The Typepad Team
Apr 8, 2010