The latest release is a 70-80MB download (depending on archive type). Where are you getting 300MB from? The uncompressed size?

If so, I must have been out of date. I just checked, and the archive is 635MB uncompressed. The joke continues.

It's amazing that people on StackOverflow suggest using this package for things as simple as sort functions.

Boost is not a package. It is a collection of packages.

If you're using any kind of sane package manager for your builds - NuGet on Win32, distro-specific package managers on Unix etc - that's exactly how it is presented. So you don't take a dependency on Boost - you take a dependency on boost-filesystem, or boost-regex, or boost-date_time etc.

The download isn't even remotely that big. If you are talking about the compiled size, you only need to build what you use (look up the --with-<library> and --without-<library> switchs).

You can also use the included bcp utility [1] to extract a subset of boost.

[1] http://www.boost.org/doc/libs/1_64_0/tools/bcp/doc/html/inde...

300 MB is nothing today and I would guess that almost all of us can download it in less than a minute, so where is the problem?

Almost everyone who happens to live in Silicon Valley, sure. Others are not so lucky and have 1mb/s download speed.

It's not only the technical size that's the problem. It's the phycological complexity of the software that strays people away from it, and the filesize is an indicator for that. It tells me that the developers aren't opposed to add any feature they feel like, creating overengineered and bloated software.

Do most people have download speeds greater than 5MB/s?

Isn't the download speed of a typical cable internet connection in the tens of Mbps?

On the east coast I pay about $70 per month for about 50mbps.

I would be interested to here if suburbs elsewhere in the country have it much worse (or much better).

300KB/s here (NYC).

Cable? What ISP? Similar price?