For small things, I pick vendors based on a few key criteria, the biggest one is how easy it is to find what I want on their web site. (The other factors are delivery time and if they’ll take a PO without needing any other information from me.)

McMaster-Carr has the best web site I’ve ever used (pictured above). It’s fast and intuitive to find what I’m looking for, and the information I need in order to select the right product is right there on the product pages (I don’t have to click through to a PDF or run a Google search to remind myself which type of aluminum is the lightest). It’s way ahead of the competition. Aesthetically, their all B&W line drawing look is pleasing, but it would be nice if I could get a color photo at some point in the search. Thier search interface works so well that sometimes I just use it to find a manufacturer item code, which I then use to find something on a different company’s web site.

Digikey recently released their Dynamic Catalog and it’s wildly underwhelming. As far as I can tell, it’s simply a reformatting of a portion of their offerings. Their basic search system (pictured below) is showing its age, but it’s decent.

Mouser’s web site is functionally similar (below).

Sparkfun’s web site is well done. They have a relatively limited catalog, so it’s easier for them to keep things nice without needing a complex search interface. What’s really nice is the discussion thread on each item’s page.

Thorlabs’ web site deserves high praise as well.

Related post: Catalogs as Textbooks (popular)

The Labrigger editorial team was a little disappointed in the consistently excellent xkcd when the recently posted list of approximations included:
Number of seconds in a year = 75^4 (accurate to 1 part in 400)
and
Number of seconds in a year = 525600 x 60 (accurate to 1 part in 1400)

In our physics classes, we always used:
Number of seconds in a year = Pi x 10^7 (accurate to 1 part in 200)
It’s still very accurate. And most importantly, the factor of Pi makes for some fascinating cancellations in calculations.

Nature has a special issue on the future of PhDs. It’s no secret that unlike the US MD program, PhD programs are not regulated and produce way more PhDs than there are jobs for PhDs.

The articles in the issue offer some suggestions. Many seem straightforward and pragmatic, such as studying the demand for PhDs in the workforce, and then changing PhD program sizes and training to match that demand, rather than the needs of academic laboratories.

I’d like to see this idea get some sort of momentum. We need a Flexner Report for PhDs. It won’t be the same, of course. For many reasons including the fact that medicine is arguably a more stereotyped practice than the varied work that PhDs do. However, the MD system is infinitely better regulated than the PhD system and there’s certainly a thing or two to be learned from it.

Perhaps these issues will be addressed on a field-by-field basis. For example, the prospects for economics PhDs are actually pretty good. They have a decent system in place. Why doesn’t this happen in other disciplines? Here’s one opinion.

The articles in the special issue of Nature offer some interesting statistics, such as the growth of PhDs granted across different countries.

BTW, here’s the annual population growth for some of those countries, for reference. They’re all right around 0-1% for the time period covered above. (source)