I’ve been pretty quiet lately, other than short posts on G+, because I’ve been grinding hard on NTPsec. We’re coming up on a 1.0 release and, although things are going very well technically, it’s been a shit-ton of work.
One consequence is the NTPsec Project Blog. My first major post there expands on some of the things I’ve written here about stripping crap out of the NTP codebase.
Expect future posts on spinoff tools, the NTPsec test farm, and the prospects for moving NTPsec out of C, probably about one a week. I have a couple of these in draft already.
I miss the general tech and political posts. We’re in the midst of a huge and fascinating political realignment caused by Trump’s election. I’d love to hear your take on it, and that of your other regulars.
Greetings from Canada
@PapayaSF, seconded. Missing the political musings.
So many interesting topics like “Trump”, “Popular votes” , “States rights” , “South vs North”.
It isn’t exactly the same as Eric’s magical/mythical analysis, but the Secret Sun blog has been doing a pretty good job of covering current events and politics from a synchromystic point of view.
I agree with PapayaSF and Bumasankar. Personally, I also yearn for posts about Zola. :-)
Anyway, are you implying that you’ll have more free time once version 1.0 is released?
>Anyway, are you implying that you’ll have more free time once version 1.0 is released?
It seems likely.
Zola will be making a special guest appearance in a near-future post on the NTPsec blog
Agree with the sentiments above, but Eric is doing the world a solid by fixing an essential IT infrastructure operability tool. Our sacrifice is everyone’s gain. Thanks for all you’re doing.
About NTPsec blog:
* why there is a “Sample Page” link in top right corner?
* a blog without comments is not a blog[1][2]… though it is hard to have static-generated website and dynamic in nature comments
[1]: https://blog.codinghorror.com/a-blog-without-comments-is-not-a-blog/
[2]: https://blog.codinghorror.com/please-read-the-comments/
>a blog without comments is not a blog
We’re looking into bringing up Disqus comments.
In ntp.conf you say “keyboard” where I think you mean “keyword”.
This appears in the snippet in Stratum-1-Microserver-Howto:
# Note: in theory we could use the “pool” keyboard here; it’s
# buggy in NTP Classic, but not in NTPsec.
(Aside: I would have reported this via gitlab if there was a way to do that without the ceremony of signing up for yet another account…)
TomA:
You’re right. But, to paraphrase Homer Simpson: “I don’t care about IT. I care about M-E – My Enjoyment.” :-P
esr:
That’s great! (You missed the period, though.) I also hope the second Beast video gets released eventually, for you said it will feature some feline footage.
Be aware that it appears to have some issues. :-(
>Be aware that it appears to have some issues
Fully aware, but could not identify a batter alternative.
I’m doing some of that with a similar bent (though I’m kinder to Clinton, who made a blindingly obvious in retrospect strategic error, and got blindsided), over at reconstitutedcoprolites.com.
I lost a bet and have to maintain a politics blog.
Troy Hunt (security researcher, creator of haveibeenpwned) also moved to Disqus, when moving his blogging platform to hosted Ghost. It looks, from what I remember, that security is solvable (after Disqus fixed few things).
> … the prospects for moving NTPsec out of C …
To what: Python? Rust? Go? (the last two being compiled languages).
>To what: Python? Rust? Go?
Rust and Go are our leading candidates.
Happy birthday, ESR. :-)
@PapayaSF thirded. If any of the regulars here have another blog where this political stuff is discussed, let us all know so we can join there!
I disagree that Clinton’s strategic error was blindingly obvious. She was caught between what her party’s new core base wanted (extreme Progressive policies, identity politics/BLM) and the economic and social concerns of Midwestern white voters who used to be the core of the Democratic party, but no longer are.
It’s fascinating to see that Trump drew essentially the same share of black and Hispanic voters that Romney did four years ago. Granted that it was low, it didn’t crash to zero despite the rhetoric. Journalist interviews with black voters post-election suggest that they are very writing to give Trump a chance. Now that Trump had backed off from building the giant wall and focusing on trade, is give him a fair shot at getting Hispanic and black support for Republicans in four years.
Meanwhile the Democrats are doubling down on their failed divide and conquer identity politics. They complain that the electoral college cost them presidential victory, ignoring that they did very poorly in House, Senate, and gubernatorial races.
2016 is going to go down into history as one of those historic breakpoints like 1914, 1932, 1945, and 1964. That’s guaranteed by the one-two punch of Brexit and Trump. It’s going to be fascinating to watch European elections over the next couple of years. Le Pen is almost sure to make it to the final round.
Cathy
C is the simplest available portable compiled language that I know of; is there another? If there is a simpler low level language, that compiles to LLVM AST representation, then it could work as a substitute. Garbage collection is a nice feature, but not for something time sensitive.
>Garbage collection is a nice feature, but not for something time sensitive.
The Go people have an interesting story there. First, the language runtime lets you turn off GC in critical regions. Second, they have engineered the garbage collector carefully to hold the duration of stop-the-word pauses down to on the close order of a millisecond. The combination of these is probably good enough for NTP’s needs – ntpd is not as real-time-sensitive as you might at first assume.
Since I can’t really ask there, I’ll try here.
It looks like that 99% of the heavy lifting is done by the clockmaker script. Do I need to run the tests by hand or will the clockmaker script do that before I move onto the next step?
Does it need to go to the outside to get a time check or can it just use the GPS to get the time? I’d like to put it with all my other Pi servers, but they are behind a firewall that does not let them send out to the internet.
Thanks!
>Do I need to run the tests by hand or will the clockmaker script do that before I move onto the next step?
You need to run the tests yourself. Making clockmaker interpret test output would have been a bridge too far.
>Does it need to go to the outside to get a time check or can it just use the GPS to get the time?
It needs check servers, because your GPS won’t always have signal.
“Rust and Go are our leading candidates.”
Allow me to lobby against Go due to its developers’ total acceptance of SJW-style political correctness. Their group has bought into Coraline Ada Ehmke’s ode off conduct hook, line, sinker, and bass boat-towing Suburban.
Reluctantly agreed with Jay. I’d heard lots of interesting stories about Go and was finally looking at jumping in with some IoT project or something when I saw the express handbasket to Nopeville. I don’t have the emotional budget for the morass that that sort of nonsense stems from and breeds.
> Allow me to lobby against Go due to its developers’ total acceptance of SJW-style political correctness.
By that standard, all three of Go, Rust, and Swift are out because they all have SJW-style codes of conduct. Swift, in fact, actually has Ehmke’s monstrous Contributor Covenant. (Go does not; its CoC is based off of Django’s. It’s still pretty horrible, though.) Really, these projects all need to be forked to get them away from the SJWs. (And LLVM too, which Rust and Swift use, while we’re at it. They’re also looking at adopting an SJW-style code of conduct. Also, Tanya Lattner, the creator’s wife, has gone full-on SJW.)
I’ve been thinking about starting such a fork, but (as you can see) I happen to work for a company that has a significant interest in both Go and LLVM, so I doubt they’ll let me do anything of the sort–especially if I say it’s because of the codes of conduct.
@Jay Maynard, @Christopher Smith –
> Allow me to lobby against Go due to its developers’ total acceptance of SJW-style political correctness.
Yeahbut…..
Isn’t the creed of the true hacker “Show Me The Code!”?? Let’s assume the development of Go has become a festering nest of SJW cronyism and jerk-off cod-Marxism. But, let us also assume the language itself is robust, freely available, and Does. The. Job.
Should we give up on the good tool just because the toolmakers are fools and/or idiots?
Wouldn’t it be similar to trying to boycott Mozilla over Eich’s personal values? Ya know, exactly what the SJW camp tried to do. I’d personally vote for judging Go on its own merit.
Disclaimer: I have a bias for Rust :-)
John D. Bell:
Exactly.
Our host, it seems, discarded Firefox in favor of Chromium/Chrome over the Eich incident. I’m not willing to do that, for I find Firefox superior to the alternatives: for one thing, the other graphical browsers I’ve tried don’t allow you to highlight text discontinuously. I’m also quite attached to NoScript, and there are a couple of other reasons as well.
Admittedly, WebKit-/Blink-based browsers handle emoji better than Firefox (at least under Ubuntu and Ubuntu MATE). But it’s not Mozilla’s fault that those pointless things have become so popular among websmiths. Perhaps they should be included in the HTML Hell Page. ;-)
Anyway, the point is: since I find the alternatives markedly inferior, I wouldn’t boycott Mozilla even if it were run by the Devil himself. And I certainly wouldn’t seek refuge in a Google product, for that company’s not without its ethical issues either.
Jay Maynard and Christopher Smith:
Incidentally, Rust is sponsored by Mozilla. Does that make it undesirable? If Rust and Go both have links to SJWs, the choice between them will have to be based on some other criterion.
> C is the simplest available portable compiled language that I know of; is there another?
One that I like, but that hardly anyone seems to have heard of, is D. I ran across it while looking for a language that 1. compiled to native code, 2. had native string and dictionary types, and 3. did automatic memory management and garbage collection by default.
The linux toolchain for it at the time was crap, as I recall, which eventually led me to drop it. I don’t know if that’s still the case. I hope it isn’t.
If the maintainers of a project show every intention of rejecting sensible contributions, then for a developer who’s experienced enough to be offering useful ones, that should be a consideration. Outright forks usually aren’t economical.
> If the maintainers of a project show every intention of rejecting sensible contributions, then for a developer who’s experienced enough to be offering useful ones, that should be a consideration. Outright forks usually aren’t economical.
Well then, what can we do? The way I see it, at this juncture, we have three choices:
We can avoid them outright. But, as John D. Bell points out above, we’d then be missing out on whatever features they’re offering.
We can put up with their SJW-infected politics. But I imagine that’s not really acceptable to any of us.
Or, we can try to salvage the tools from the toolmakers by forking the projects. (Alternatively, we could use the tools without engaging in their surrounding communities, but somehow, I don’t think that’s tenable in the long run. Eventually, their obsession with “social justice” will ruin the projects themselves.)
I don’t know about you, but at this point, I’m ready and willing to fork all these projects and save their code bases from SJW subversion.
Is there any indication that Go maintainers/authors’ propensity towards SJW-ness actually affects the project? Will they unreasonably reject contributions because of politics rather than technical merit? Again, Brandon Eich’s (formerly secret) support of Prop 8 never affected his business and colleague relationships; whatever views he held in the larger world, he practiced “shut up and hack” where it matters.
For many initial years of its life, the D language and compiler was proprietary. The author even made threat against making alternative implementations. This pretty effectively neutered its chance at becoming significant in the critical early stages of its life, I doubt it has any chance even after becoming open. It also largely tries to solve the same sort of problems that Go tries to solve (mainly being an alternative to C++/Java/C#), while maintaining a large bulk of the bloat that C++ has. It’s interesting, but… not really all that clear why or how it is desirable.
You should consider D, very nice language. Regarding a CoC, its creator is against them, a minor point in its favor:
No. People who need to be told what decent behavior is won’t pay attention to such a document.
http://forum.dlang.org/post/mvjl43$129p$1@digitalmars.com
I did not mean that absence of a Code of Conduct is license to abuse others. Just that a CoC is itself insulting, paternalistic, and not a solution.
http://forum.dlang.org/post/n0r123$1s8r$1@digitalmars.com
@Cathy
> Le Pen is almost sure to make it to the final round.
I really hope that it won’t happen. I’ll be ashamed. The story is not yet written. I see a lot of comparisons with Trump in the press and I think that this is wrong. Even if she tries to be associated with Trump in the minds of the French. She IS a racist. Moreover people from FN are incompetent at many levels. In France, if you are not a racist there are no reason to vote for Le Pen. And you can still vote for nobody. Or just do not vote. Unfortunately, there are people like that (if you can understand).
Le Pen is quite discreet currently. Especially since the victory of François Fillon. Fillon is able to gather some voters from “extreme right”. Even if this is not an easy task. He has good cards in his hands. The party is already in order (unlike the socialists) and Sarkozy was pretty good at that.
The candidacy of Valls is also a good news. Socialists are not yet in order and are lagging behind. They must solve their internal problems quickly (they are very disappointing on this point, as usual). In France, Valls is the symbol of a more modern socialism (not comparable to Macron) and he may be able to get some voters from the “middle” or “republicans”. Perhaps those who fear Fillon’s liberalism (I do not like comparisons with Thatcher) or those who do not accept some kind of rapprochement with “ideas” from the extreme right, in order to get more votes.
@swanson: I guess that explains why few have heard of it. Kind of a shame. I found it pleasant to work with, but admittedly that’s by comparison to exactly the languages you suggest it’s trying to replace.
> I miss the general tech and political posts.
Voting for tech and against politics for a while. I’ve been watching Eris take a dump on every outlet I read for so long that I forget what useful things look like. I’m getting kind of sick of the Two Years Hate.
I prefer Go to Rust. Rust is like someone chopped bits out of C++, Haskell, and Erlang. Then welded them together. Each language has its own high cognitive load. Combining them makes that cognitive load go up exponentially. Go on the other hand, is sweet, like if someone took the good ideas from Python and C, and then made it even simpler.
>Rust is like someone chopped bits out of C++, Haskell, and Erlang. Then welded them together.
You put your finger on one of my concerns here. I haven’t coded in Rust yet, but I’ve read a lot of its documentation and the welding seams in the language design are pretty visible. You are correct that Go seems more of one piece.
Mycroft Jones:
Heh. In the comment thread for “When hackers grow old”, Joshua Brulé wrote:
To which Eric replied:
“Weird sex”. That’s funny. As if sex could be non-weird!
New post on the NTPsec blog:
https://blog.ntpsec.org/2016/12/07/down-on-the-test-farm.html
Mrf. Is a non-SJW fork of Go likely to gain the sort of mindshare that Go has? Forking is the open source world’s way of routing around brokenness, to be sure, and CoC are one form of such brokenness The concern is that they will run off good contributors for not being politically correct in all their doings, thereby rendering their product less than it should be.
For a non-SJW fork of Go, or some other open source project, to succeed, that implies that it would need to have a base of contributors that have been rejected, or scared away, but the SJWs, and those contributors would be capable of improving the product. I’m convinced there are people like that. The question is, can they be attracted to a fork and can they provide enough visible improvement to attract users away form the original?
I’d really love to see a fork of some major SJW-infested project take off as an existence proof that a CoC isn’t needed, even now in these days of rampant political correctness, to produce a technically superior body of work. I don’t know what the answer is to which project needs it that satisfies the criteria I laid down above, but I’m certain there is one. Just need to identify it.
Interesting article about using clocks for ruling out exotic sorts of dark matter (and from scientists from my corner of the world!)
http://arstechnica.com/science/2016/12/existing-atomic-clocks-could-help-test-for-exotic-form-of-dark-matter/
About NTPsec: would it be good idea to try to apply to OSS-Fuzz project?
https://github.com/google/oss-fuzz/
https://lwn.net/SubscriberLink/710534/44ef6221276d9ba2/
https://lwn.net/Articles/710534/
>About NTPsec: would it be good idea to try to apply to OSS-Fuzz project?
We already have.