Discriminating indiscriminately...

Wednesday, 5 March 2008

Music trends

Last.fm is great site for finding out more about music, for recording your own music preferences, and potentially connecting to others with similar preferences. I've not used the latter function, although it is interesting to see how many people out there have eerily similar music preferences.

Interestingly, I see Google has tentatively moved into the same field with Music Trends, a Labs project. It's nowhere near as fully featured as Last.fm, but certainly one can see the potential there. Similarly, they use a local client (in this case Gtalk) to snoop on a music player to upload the playstream to the server.

The links from the Trends page lead to an artist/song customised search result page. I had rather hoped to see a Trends-like chart showing temporal popularity, but I suppose they're not implementing the infosthetics fodder first and are concentrating on the basic features. Certainly looks like the monetisation is not going to be an afterthought, here!

For now, last.fm is safe since it offers song previews and you can usefully tag songs. Usefully, last.fm artist/song pages often rate highly in search results too. Perhaps music is too specialised to have a generic solution be worth using!

Sunday, 24 February 2008

Those precious seconds

I've been using a ThinkPad T60 at work for the last year and a half, and when I first got it, I thought it was fantastic -- very responsive, hardly ever crashed, and the battery life is quite good.

However, I now find that startup -- whether resuming from standby or cold booting -- is an interminably long process. This is probably partly because NTFS gradually degrades due to fragmentation (something that defragging now and then doesn't really seem to affect in a perceptible way), but the reasons seem to be different depending upon whether it's bootstrapping or resuming (more posts than you can shake a stick at here).

When bootstrapping, the execution of multiple applications seems to kill performance, mainly due to disk access requirements. Whether this is due to multiple processes trying to load libraries at the same time, bad virtual memory implementation (perhaps also a side effect of the massive paging file being easily subject to fragmentation), or no prioritisation between background app loading to guarantee UI responsiveness, I don't know, and frankly don't care. A good analysis is here on the ThinkPad forums, including recommendations for removing various superfluous ThinkVantage startup software items.

When resuming, I've found the biggest improvement was to disable the anti-virus service that came pre-installed by the company IT department (in this case, Trend Micro OfficeScan). This service seems to cause at least five minutes of disk thrashing as it does some sort of resume-check. It could be that this is due to the interaction of the scanner with some other software, like Skype, but disabling the service does mitigate the problem to some extent.

Perhaps this is due to Microsoft not coordinating with software vendors to really test OS responsiveness under real-world conditions: namely, having multiple programs installed and frequently sleeping and resuming. One shouldn't have to dig into the registry to disable prefetch as a matter of course (it didn't work for me anyway).

The sad thing is that my old Apple Cube (literally a museum piece), with a 1.7GHz PowerPC CPU accelerator installed, seems faster and more responsive than my T60. Good news though, James Fallows is finding the MacBook Air a more than adequate replacement, and some former MS junkies now catalogue the reasons to switch to a Mac!

Saturday, 23 February 2008

Sharing ideas efficiently

I like sharing ideas and news that I find interesting with a few others who might be similarly interested. A long time ago this was just by email but for a while I've been using del.icio.us both to record bookmarks so I can find them quickly later, and to share them with a few friends who have added me to their network.

More recently I've taken to using Google Reader as the most efficient way to get my fix of info from sources I trust to have a high S/N ratio. This has the added benefit that I can mark items for sharing which are automatically published to what looks like a pseudo-blog, and can see the items my Reader-using friends have shared for me. End result: higher quality information, if my friends and I are roughly interested in the same sorts of things, or quality of information (very subjective, yes). The final result was that I started this blog since sometimes, I can't find a good post/article by someone else -- trust me, if I could, I'd link to that instead to save time.

However, at work I use Skype a lot for communication. Skype multi-chats are fantastic ways to organise cliques of information, as impromptu chats on a particular topic can be created in seconds, and they can be persisted as long as useful. Whilst I use the Skype mood message to publicise a link now and then, I find dropping a link into a multi-chat a great way to rapidly share information.

The problem with these different media channels is obviously that my Skype-using friends don't necessarily see my other shared items (unless they also use Reader etc), and my non-Skype-using friends don't see my Skype-posted links. Either I'm going to have to convert one group to use the channel of the other, or double-post everything.

Perhaps this means Skype needs a way to publish links from chats to a feed, or publish feed contents into a chat. No doubt, this can done with a plug-in, but is it worth doing?

Friday, 22 February 2008

Geeks and schmoozers

Since I've relatively recently transitioned from an engineering-focused career into a product management role, I'm interested in the interface between purely technical skills -- such as those excelled in by "geeks" -- and communications skills -- yes, that's what schmoozers do.

This article in the FT ("CDO buffs who schmooze could resolve a financial mess") shows that the distinction between communicators and technicians is a tension that exists in other thought-based industries too. I've been working on my schmoozing skills a lot: if anything, a product manager has to do loads of schmoozing since nobody actually reports to them, and therefore to get any work done, coercion of engineers (or financial product designers) is not an option. Some other product managers I have worked with have not got the technical background but are great communicators, and have to work hard to understand the technical nature of a problem in order to understand how to make a decision.

Which is more effective, a geek who studies communication skills to transmit their message to a wider audience, or a schmoozer who somehow picks up enough understanding of a domain to not only gain the respect of the local geeks, but also can make better decisions than them?

In my experience I've seen more of the former than the latter. A schmoozer who knows how to use geeks is common enough, but I haven't seen many examples of a person with a predominantly communication-oriented mind manage to pick up an easy facility with abstract concepts and models. Good MBA courses are supposed to be able to take a marketing grad and churn out an investment banker, but I'm a little sceptical of this (although I would love to have my scepticism proven wrong!).

Of course, all this is a simplistic view of what Simon Baron-Cohen has been working on: the tendency for a brain to be systematising or empathising (EQ SQ theory). Whilst he noted that there are statistical tendencies for men to be more systematising and women to be more empathising (something James Watson has been slated for supporting, judge him for yourself), it seems to me that many men tend to be heavy empathisers ("I just never understood math") and plenty of women are strong systematisers (Hillary Clinton springs to mind).

Will there be a simple way to understand and identify why this is in the future? Presumably brain wiring is a phenotype and the underlying genetics can be determined one day; will optimum training techniques for those with a perceived deficit in one quotient be developed? I think so.