jack12

OSH Park (formerly known as Laen's DorkbotPDX group PCB order) is very popular for prototypes in the US maker community (look for the trademark purple PCBs), and at $5/inch^2 with free US shipping it's cheaper than batchPCB for smaller boards, in addition to getting 3 copies of the board for that price. It's very high quality and it's all done at a US fab house with the 2-layer batches sent out every 2-3 days and then shipped US Mail, so if you're in the US it's pretty fast for the price.

And for pure price, Seeed and ITead (both are rumoured to use the same PCB fab) are very popular with their $10 for 10 copies of a 5cm*5cm board service. The quality is reportedly much more variable with these guys, and the shipping takes a lot longer to reach the US, but you can't beat the price. For the rest of the world, the shipping from China is probably cheaper and faster than from OSH Park, so unless you need the quality of OSH Park this is probably both cheaper and faster.

unbound is a resolver, which I think is a bit different from this. unbound's NLnet Labs also make NSD, which is their authoritative server, and that is one of the servers mentioned and compared to on YADIFA's site.

Are you sure you're not describing cross-platform with Windows Phone instead of Windows 8?

I thought for Windows 8, Microsoft is pushing WinRT as the cross-platform solution, and that "binaries" crossing from x86/x64 to arm will work as-is for javascript, as-is (if pure) for CLR, and with no modifications but requiring recompilation for native (C/C++) WinRT apps.

Oh, nice work! There really isn't much point in running an indicator if it's just doing a DNS lookup to _guess_ if you were using IPv6, it covers up mistakes too easily. So it's nice to see it done properly, and the way you show the protocols of additional hosts is really slick too.

The indicators all just being AAAA or not instead of IPv6 or not was my main complaint last time I tried what was available on Firefox, but your comment lead me to the Firefox clone of IPvFoo: IPvFox. Thanks!

The only description of the service in that article comes from the publishers' complaint, and makes it sound like Boundless provide little more than a photocopy of commercial textbooks run through a ransom-note filter.

Yet "When asked to describe how his company pulls together the open-education content to produce its digital textbooks, Mr. Diaz declined to elaborate on the process" and so the publishers ended getting to establish the first and only concept of the company to a huge audience. And it's not like people can check out the website for themselves, since it's in closed beta.

Staying quiet here really doesn't seem to have been the best strategy.

The NYTimes article says that the initial offering of BATS "would have been the first company to be listed as well as traded on the exchange", and the post mortem currently at the top of http://www.batstrading.com/alerts/ (no direct link that I can see) says that the A-BFZZZ shard "encountered a software bug related to IPO auctions".

I read that to mean that this is the first time they've ever used their IPO-related code in production and it failed badly. So it sounds like it was the normal sort of bug, in their own code, and their own IPO directly triggered the bug.

Arch has the fewest deviations from "regular" linux of any distribution I've tried, probably because keeping everything as close as possible to vanilla is one of the core principles of the distribution. It still has all the convenience of the one command software install or upgrade, but it doesn't place itself between you and the software once it is installed like so many other distributions do. It's great if vanilla linux is what you've grown up using, it's probably a moot point if you don't default to trying to do things the standard, vanilla way.

I have had other distributions autogenerate and overwrite hand-modified /etc files, and not just system-level ones but even daemon config files! It was something like the distribution required you to edit a distribution-unique "local changes" file for the daemon, not the daemon's actual config file, all so the package manager could incorporate parameters out of some sort of "friendly" config database into a new autoregenerated config file.

Many distributions require that you run a custom command just to do something as stupidly simple as create the /etc/localtime symlink, sometimes even dropping you into a GUI (to create a symlink! Arch is guilty of autocreating this symlink too, but at least it's controlled by a single ASCII line in the distribution's one main distribution-specific config file, rc.conf). Or some distributions have "SysV"-type init systems but require you to use a custom command to handle creating and deleting the symlinks for starting and stopping daemons. Again, usually to allow a package manager to alter those same files without having to figure out what a human has done to them. Those are the sort of "hoops" I'm thinking of.

My complaint's not that those systems exist for people to use, especially if it makes things easier for them, it's just that everyone is forced to use them. Those custom commands/the package manager integration is usually pretty delicate and almost always breaks completely if you try to do things the standard way, without going through them, or if they don't break they stay resilient by silently erasing your changes.

I'm sure there are great advantages to be had by committing fully to a given distribution's custom systems. People managing large or many systems probably find a lot of problems solved by the distributions' custom methods of doing things and the tight integration of those methods with their package managers (but I'm just running a laptop). I'm sure they're also just fine if you first learn how to do a specific configuration change on that distribution, and then don't even need to care what the real output of your manipulations of the distribution's configuration system is.

But if you already know the form of the change you want to happen, the exact line in a config file that you want changed or the exact flag passed to a daemon on startup, it's incredibly frustrating. You have to wade through all of a distribution's weird changes to a piece of software and its added config system layers just to figure out how to mangle your change in such a way that the distribution will unmangle it back into the exact form you could have just edited into the program's config file by hand in 3 seconds.

It's been a long time since I've tried Debian, I don't know what it's like in its modern form, how much you are required to actually cooperate with the distribution's unique systems or if you can ignore them without breaking anything. But the presence of debconf, alternatives, and update-rc.d all sound very discouraging.

I'm not sure exactly when I moved over to Arch from Crux, I'm sure the distribution was at least a few months old by then, but it's been a great distribution as long as I've known it.

It does suffer the occasional dip into making things more complicated than they need to be, some element of the distribution straying towards the "SysV" darkside and away from the "BSD" ideal, or an upstream's configuration system just getting too insane to work around any longer, but everything always works back to some sense of balance after a few years. And it does feel like upstreams come around to the Arch way of thinking more often that the other way around.

I do wish it was a bit easier to automatically build or just download pre-made packages with some of the more popular compile flag variants, ports- or portage-style, but not so much that it casts a shadow on all the other benefits of Arch.

Still, it would be nice to be able to install a console vim with python and ruby interpreter support without having to install gnome too, or being able to install java on a headless server without having to stick half of X11 on there, or not having to install Apache because you want to change nginx's modules. It's always possible to change the PKGBUILD and recompile, but it seems like just changing one enable/disable switch to ./configure should be easier than it often is (vim is especially a pain to keep a custom PKGBUILD of, the ABS PKGBUILD seems to undergo massive revisions every few months).

But really the distribution is just great, the best out there. It lets you use pure Linux and avoid all the hoop jumping other distributions make you do in order to keep their package managers happy, while still giving you all the benefits you want from a distribution like automatic pre-compiled upgrades. And that's enough to make it the best balance of a distribution out there for me.

I don't get the Post's headline. EA published 'SimCity Societies' in 2007 and an expansion pack in 2008, they even mention it in the article. With the inevitable impact of The Sims' mass appeal, and the industry-wide interest in broadening their market by making games, well, simpler, perhaps the article writer feels that nothing has been a "real" SimCity game since Sim City 4, I guess?

It's an important point, whether the series is returning to it's roots of complexity from ten years ago, or if it will be a continuation of the franchise's modern-day releases. But the article leaves a very confusing impression.

I suppose they might mean to highlight this is the first SimCity game developed by Maxis themselves since 2003, but with Will Wright gone and 10 years of staff turnover, is there really much difference between a SimCity game being developed by "Maxis" instead of by any other outside developer?

There's a newer version (1.3.4 vs. 1.0.6/1.1rc2) at http://pencil.evolus.vn/en-US/Downloads/Application.aspx

I don't know if it fixes any of the problems deyan mentions (or which version he used), but 1.3 is about 2 years newer than the version on their older site. They really should delete the old copy of their website's URLs, and redirect to their new site.

like everyone else said, the link's not about the UI. But on that subject, autocutsel (http://www.nongnu.org/autocutsel/) can keep all of your X11 clipboards synchronized, so that anything copied to the clipboard can be MMB pasted, and anything selected with the LMB can be Ctrl-V pasted.

Me too, but on PuTTY Tray's download page:

"Last update: November 25th, 2007.

Please note: I have decided to stop development of the PuTTY Tray patch, and I have no plans to resume it at any point in the future."

The title and text, etc. are all directly responsive to the linked contemporary WHT thread where someone asked about a 64mb plan on this specific service, and if it would be enough for 18 static sites. The WHT commentators all said no.

I'd be inclined to believe it is not spam. Mainly because sites designed as hosts for ad-articles/spam don't tend to build up communities or stick around this long (this article is over 2 years old).

But also because the site covers multiple providers, not just this one, sometimes favourably, sometimes not, always without any real benefits to the owner. It does provide a (presumably no longer valid) referral code for the particular service discussed in this one article, but it's an extra click away from this page, and the site goes to very large pains to point out it is a referral code: http://www.lowendbox.com/blog/vpslink-february-sale-and-10-o...

Also, to my knowledge, that is the only time this site has ever used any sort of referral code, and it was fully highlighted and explained. It also wasn't an affiliate code, it could only have been used to cover the site's hosting fees with a service the site owner had been using for the last 5 years.

Basically, it doesn't smell like spam to me. But, I have been enjoying the site for a number of months before it appeared here, so I'm less suspicious.

Yes, but beyond the headline is the article :)

It actually ends with MySQL + Nginx + PHP + WordPress all running in addition to the 18 static sites, which does seem a bit unlikely to hold up very well, even if the setup shown does leave the VPS with 48MB free.

Worth pointing out that the site behind this is itself running Wordpress on a 80MB VPS (plus swap): http://www.lowendbox.com/blog/quickweb-7-128mb-xen-vps-in-la...

That's an old article, but LowEndBox is a fun site for checking out the bottom end of the VPS market, these days it's mostly only postings of very cheap providers ($7/month limit) the blogger has been sent or seen on WebHostingTalk. Then the ensuing discussion of how long until the company behind it goes bankrupt. 64MB VPSes like this don't really get offered anymore, while 128MB and 256MB offers are much more popular, usually in the $15-$20/year range.

HE also offer a free DNS service now, considering you've already made an account with them.

Being able to turn on and off features based on login or randomness seems like a pretty common feature when you get big enough for it to be meaningful to do tests on portions of your userbase, or have multiple tests running at once.

Disqus have open-sourced a similar system for Django-based sites, Gargoyle: http://code.disqus.com/code/projects/gargoyle.html

You basically stick if(abc-is-active)s throughout your code, and then it has a fancy frontend for defining when abc should be active, based on user, ip, randomness, etc. They showed it off in their PyCon 2011 presentation, at about the 10:00 mark: http://python.mirocommunity.org/video/4256/pycon-2011-disqus...