Re: Oh My!
John Wyndham's 'The Death of Grass'
That would be John Christopher (alias Sam Youd). But it's not difficult to confuse the two, since Wyndham did have several notable entries in the post-apocalyptic genre.
32 publicly visible posts • joined 27 Oct 2008
Heh. I've recently handled Serbian currency, so this column was a serendipitous reminder: the 2000 RSD banknote also has a scientist on the obverse, Milutin Milanković оf the ice age cycles fame. It's a rather strange portrait, from the mid to late 1920's in my estimation, where Mr Milanković faintly resembles the emcee from Cabaret.
I thought that Firefox had gone whole-hog with a Rust engine.
Not whole-hog, which would be bad engineering, but several parts of the browser engine have been replaced with components written in Rust, the largest being the style system. There is a nice overview in the slides accompanying a talk by one of Mozilla engineers. Briefly: FF is 9M lines of C/C++, 160K of which was the old style system, now replaced by an 85K-line rewrite in Rust.
More Rust components will appear in the future.
The developed/implemented confusion is the most likely explanation, indeed. Furthermore, shame on El Reg for not checking the history more thoroughly: RFC 2460 is not the first specification of IPv6, RFC 1883 is. That specification is from 1995, making a working implementation "in the 1990s" a more realistic proposition.
The blog post which announced the patching overhaul, with the juicily Doublespeak title "Further simplifying servicing models for W7/8.1", stated that a separate security-only patch set would be available each month, although not via Windows Update, but the semi-obscure Microsoft Update Catalog, then and now a throwback to the early noughties which requires IE and ActiveX (yes, that ActiveX) for what is basically a glorified FTP server.
The "and now" part is pertinent: the same blog post promised that the ActiveX requirement will be removed "soon". Two months later, it still isn't: try visiting the site in anything but IE, and it will redirect you to a page saying that IE is required. Er, Microsoft? Are you trying to make this deliberately difficult for those who don't want to be force-fed the whole opaque load of unrelated crap? (The question is rhetorical, obviously.)
What's even more astounding is that the non-IE promise is not even new: here's an earlier incarnation (scroll down a couple of paragraphs), plus a workaround for non-IE browsers. The "screw you, we're Microsoft" spirit is as strong as ever over at Redmond.
"the geniuses at the IETF (Internet Engineering Task Force) deciding not to make it backwards compatible with IPv4."
This reflexive, non-thinking non-argument really needs to die. There's no way to make it backwards compatible without interposing some NAT monstrosity between the endpoints. Avoiding NAT as far as possible was an explicit design goal of IPv6, which hasn't been arrived at arbitrarily, but from bitter experience with that kludge.
What about when the client has no idea nor any way to find out what the outside-facing IP address is?
It doesn't need to find out, since the problem is self-correcting: the server will generate a new Server Cookie when the outside address changes. As long as the mapping to the external address is relatively stable, the client's interaction with the server won't be overburdened with cookie recalculation. See section 6 of the RFC for more discussion of NAT issues.
The terrorists have their own routing protocol now?!
Proudly routing your packets since 1992.
Joking aside, Cisco used to have a lot more love for IS-IS as opposed to OSPF for interior routing if one decided to forgo their own IGP spawn (EIGRP, blech.) I'm not very closely following the networking game any more, so I don't know if it's still the case.
I'm too young to have personally experienced the dethroning of the mainframe by PC proliferation, but that situation has remarkable similarities with what's described herein. Need a service? Whip out the credit card and there it is! Just like people could solve a lot of their problems with Lotus 123 and dBase...
And then those little services proliferate. You find out that their integration with the rest of your processes is nonexistent to lousy. You end up with an unholy hodgepodge of little custom somethings which are partly critical to the functioning of your business, and you don't know which parts are the critical ones. So you organize a department to integrate, control and oversee the services -- and the next priesthood is born.
I'm particularly interested in seeing how it performs on low-end hardware such as a Raspberry Pi.
One especially interesting thing they're trying to do in Servo is implementing the bulk of rendering on the GPU, using game-engine-like techniques. See, e.g., this demo. This could make even RasPI perfectly capable of nice browsing performance (modulo atrocious JS multi-tentacled ad-serving and tracking dreck served on a lot of sites, natch. But there are adblockers for that.)
"Any landing you can walk away from is a good one." Applied to IT, any disruption which leaves you with your data intact is tolerable. No data loss here (or we'd have heard about it, I'm sure.) Comic distractions of the "keyboard not found, press F1 to continue" variety seem unavoidable in any sufficiently complex setup.
I took a very quick look at rust docs. Saw this line:
let x = 5;
Whenever I see "let", I think ANSI Minimal BASIC from 1980.
Am I wrong to make this association?
In this case, you are. If you're familiar with C#, think var (Rust's let is a bit more complicated than that, but it does introduce a local variable binding whose type is inferred by the compiler.)
Play it.
Report back here in xxx months.
Did that for a while. I have two hard rules for free-to-play games:
1. Make it possible to advance by skill and/or moderate amount of luck if you choose not to spend money.
2. Don't be blatantly in-my-face with suggestions for purchasing gems/coins/whatever.
(Along with a meta-rule: I won't spend a farthing on your game.) CCS violated both and got uninstalled.
Go thou and do likewise.
We're seeing the results of thorough computerisation of all sorts of ancillary tasks surrounding software development (and not only that.) A good technical manual is a feat of organisation, and lots of people suck at organisation. In times past, there were trained professionals who could take care of such things -- technical writers, secretaries, archivists. No more: everyone's expected to do everything themselves, with the excuse that there are those excellent software tools to make it possible, but tools are useless -- or worse, seemingly useful -- in untrained hands. Cost-cutting all the way; the mask of knowledge and efficiency is wearing thin, and the truth of an average bloke's (and blokette's) muddle-headedness shows through.
It is ironic indeed that it needed one of the most proprietary companies in the world to make something good out of open standards like LDAP and Kerberos. It really goes to show how important good design and forethought are in these matters. MS did a really good job of it, so good in fact that they had created a technical (and thus commercial) monopoly
Lest anyone take this at face value -- it's a moderately subtle troll. Microsoft did make a good job of... extending and/or breaking Kerberos and LDAP in a myriad of ways, some obvious, some quite esoteric. The infamous Kerberos PAC (noted as such in the Samba documentation, btw) is just one example. Why do you think that Samba 4 took so long, even with the full protocol documentation? They had to implement and integrate their own versions of Kerberos and LDAP, since using the existing (standard compliant) implementations would mean butchering them beyond recognition.
When I was in high school, seven hundred years ago, I was pretty much the only one with a scientific calculator (an HP 12C at that!). When people asked to use it, surreptitiously putting it in hex was good ("What? Six times seven is 2A? Your calculator is broken, man.") but octal was better, because they often wouldn't notice the difference.
Cool story, but HP-12C is a financial calculator -- no alternate base modes. Moreover, even if it were an HP-16C, which did have base-N calculations and then some, a typical high schooler would be stumped with RPN logic before arriving at 6 x 7 = 2A*. So there's something fishy about that recollection, innit?
*I mean 6 ENTER 7 x, of course.
If its just altitude I would imagine it should be possible to mod an android phone app to give you the proper altitude - I believe paranoia prevents most sat navigation jobies working at altitude but the signals should still be there.
Satnav is limited on a lower level: ITAR (the same export regulations that gave us 40-bit encryption 15 years ago) limits the receivers' working range to less than 60000 ft / 1000 kt (~18 km / 1800 km/h) at the same time -- the idea being that you can't easily piggyback your ballistic missile guidance system on GPS. This is usually done in firmware. Theoretically, LOHAN should be OK, since it would reach a greater altitude, but at a lower speed. However, barometric sensors are simpler and accurate enough; there is no reason to complicate the design.
The first one looks like a correct Greek transcription, the second is valid Cyrillic for a number of languages (Bulgarian, Russian and Serbian off the top of my head), and the third is Ukrainian Cyrillic. I'll leave the presumed Japanese, Chinese and Korean transcriptions to those in the know.
Paris because it suits the topic.