Posts by Michael Wojcik 12336 publicly visible posts • joined 21 Dec 2007
I've certainly enjoyed holidays in places where there's no phone service. But different people enjoy different things. Not everyone is you.
I'm not particularly interested in "relaxing", either. I'm pretty relaxed in much of my daily life; I don't need a vacation for that. I take vacation to spend more time with my extended family and enjoy a variety of activities I don't normally have much time for. But "relaxing" is not my goal.
Closed, proprietary software is dead
Yawn. "X is dead", for whatever value of X the author dislikes, is the most feeble, threadbare claim in IT. It's the flag waved by those who have no actual argument to make.
Nothing lasts forever, but most of the things in IT which some self-appointed expert has confidently declared "dead" are still around.
Yes. And while COBOL more or less encourages things like comments (the NOTE statement may have been the first explicit provision for long-format comments in source code) and meaningful variable names, that doesn't mean developers will use them.
I don't know how many times I've had to search through multiple source files trying to figure out all the ramifications of someone's SET ws-ctrl-flag-foo-88 TO TRUE.
And COBOL written in pre-COBOL-85 style, with punctuation instead of scope-delimiters, can hide control-flow errors. As can the inconsistent semantics of PERFORM across different COBOL implementations.
The z architecture has BCD in hardware. I don't recall Power having it.
Often BCD arithmetic in COBOL programs is on items small enough that they can be represented with full accuracy in one of the native CPU types, so there's just a conversion penalty before and after a basic block of arithmetic operations. You don't need to do actual BCD arithmetic until the items get too large to fit in a 64-bit integer.
In any case, the real USP of COBOL is that it's COBOL, and there's a lot of it. We (Micro Focus1) sell a whole heaping bunch of mainframe migration because there are so many mainframe COBOL applications which are enormous and stovepiped and embody business logic that's not documented anywhere. Rewriting those applications is a minefield. It's much safer to move them unchanged, or largely unchanged, to a new platform under emulation that supports mainframe aspects like CICS / IMS / JES environments and EBCDIC and mainframe pointers. And, yes, IBM mainframe COBOL dialects.
And then, once your existing systems are running on the migration platform, you can start to modernize them. Integrate with other native or managed languages. Slap web UI front ends on. Wrap pieces as services. Scale out. Whatever.
IBM selling a port of their COBOL compiler for Linux is only a small piece of the mainframe-migration puzzle. And GNU COBOL is nice (and Bruce TIffin deserves ample respect for the huge amount of work he put into OpenCOBOL and then GNU COBOL), but again, a COBOL compiler is only one of many ingredients for migrating existing COBOL mainframe applications.
1"Micro Focus". Two words.
COBOL has been modernized.
OO COBOL? Check. Managed (CLR and JVM) COBOL? Check. Inline declarations and anonymous closures? Check. Support for popular IDEs? Don't know why those things are popular in the first place, but sure, why not. Web UIs and service interfaces? Check.
And compilers have supported lower- and mixed-case COBOL source code for ages. And free-format (so no more worrying about columns), too.
As I understand it (having read explanations from some lawyers), this sort of thing can become barratry – which can be grounds for disbarment and criminal charges – but is often permitted under US law. Thiel's subsidizing of Bollea v. Gawker is a famous example.
My understanding is that a certain amount of financing from outside parties is permitted by the court system on the principle that a well-financed party in the wrong might be able to stave off justice by making it financially infeasible to sue, and this is a mechanism for plaintiffs to assemble the required resources.
I hadn't noticed the similarities between the electric-universe and sovereign-citizen types before, but now that you mention it...
They seem to share a belief in the magical power of words, too, whether it's e.g. "vibration" for the electric-universers or "proper" names for the sovereign citizens.
There were numerous proposals for increasing the cost of sending messages, and many of them were well thought out. Adam Back's Hashcash is a well-known example. David Chaum, in an interview, cites a similar earlier proposal by Dwork and Naor; and Chaum's own work on micropayments was often mentioned as an aspect of solutions in this context.
Hashcash does not, of course, cost actual money; it's a proof-of-work scheme. There are other proposals which do apply a direct cost.
The Hashcash proposal was in 1997, I think. So that's a fair number of years back, for this industry.
Payment-for-email proposals failed for much the same reason that encrypted and signed email largely failed: email is too decentralized and the big MUA vendors and MTA operators aren't interested in pushing improvements. (OK, that's a bit unfair – none of the standards for encrypted/signed email are particularly usable. PEM was never widely adopted, PGP's PKI doesn't scale and is too difficult for ordinary users, and S/MIME uses X.509 PKI which imposes a startup barrier in terms of cost. But foot-dragging by MUA/MTA implementers certainly didn't help.)
There are certainly cases of "cyberstalking" and other criminal activity being carried out through social media, and using the relative anonymity1 of social media to impede identification, investigation, and prosecution.
Australia's mooted requirement would increase the work factor for attackers to commit these types of crimes using social media.
So would banning social media, or licensing all Internet use. And you could eliminate more crime by instituting more physical-world surveillance measures. That doesn't mean any of these are good ideas, or are appropriate trade-offs between security and liberty.
Australia's government is on an authoritarian surveillance-state high. Things will get worse there before they get better.
1Of course "relative anonymity" isn't really a meaningful term, in any precise sense. What people mean by this is better expressed in terms of differential privacy, specifically in the amount of information available from a typical social-media account at various levels of effort and within the scope of various laws.
a corrupt any law enforcement or regime WILL abuse that right
FTFY.
People are very good at rationalizing behavior which will achieve a short-term goal at the expense of ideological principles or long-term social goods. And law enforcement and other government functions are still run by people.
There are other variants of cryptocurrency. Ether is tied to the Ethereum smart-contract platform. There are various types of proof-of-stake cryptocurrencies. There are various forms of stabilized cryptocurrencies. There are NFTs, and there are other potential new types of derivatives.
So not all cryptocurrencies are equivalent to Bitcoin; some have other affordances and properties that are of interest to certain potential users.
That said, the only advantage for the central banks that I'm seeing is the potential for increased surveillance. Otherwise you'd get more benefit from just streamlining electronic transactions using your existing currency.
I find cryptocurrency interesting from the technical and economic standpoints, but I have no interest in using or investing in it. (And non-cryptocurrency use of Blockchain is generally just a poor application or the wrong choice of tool. Nagle graphs and append-only data structures are useful; Blockchain is a pretty dumb variant of both.)
Why are unions such as a problem for Amazon US ?
It's not so much that they're a problem; they're just cheaper to fight than to have, at least at the moment, in Amazon's estimation. So they'll fight them. It's all about estimated return on investment.
Because anti-union sentiment is relatively strong in the US (compared to many other countries), fighting unionization is often successful here, which raises the expected return on the investment in union-busting. If Amazon thought unionization here was more or less inevitable, it would determine that the more economical course would be to make nice.
we have clearly reached the limitations of current US copyright law
In fairness, that's basically what Breyer said. And the Roberts court really prefers narrowly-scoped decisions, which is why they left the "does copyright apply to APIs" question open.
Federal legislation to clarify how Title 17 applies to software would be the best remedy, if it's decent. But getting decent legislation written and passed is non-trivial. And this isn't one of those areas where the states have much discretion, so Congress can't rely on the usual legislative process of using the states as a laboratory, then copying what they like.
IANAL, but I'm pretty sure this wouldn't be contempt of court (in the US). It's just expressing an opinion, which is strongly protected by the First Amendment.
I can proclaim on the street-corner, or publish an editorial in the New York Times,1 excoriating various SCOTUS decisions on questions of the Fourth Amendment – an area where they have Not Done Well. That's not libel (or slander) or contempt. I can, in fact, profess literal contempt for the court;2 that's not "contempt of court" in the legal sense either.
Defying a court decision is contempt. Disrupting proceedings is contempt. Disagreeing isn't.
1Oh, you know they'd love to have it.
2Not that I'm actually contemptuous of the court. But if I were, I could. Well, I suppose I could even though I'm not, if you see what I mean.
The bar for libel and slander in the US is appropriately high, not absurdly low as it is in some other jurisdictions.
Also, of course, there's nothing libelous or slanderous in claiming that a court arrived at the wrong decision. That's a clear matter of opinion, and under US law opinion is never libel or slander.
If Oracle published a statement, that, say, accused Breyer of taking a bribe from Oracle, that would be potentially libelous (though it's such an absurd hypothetical that it's hard to say whether anyone who matters would take it seriously).
Mike Masnick has a good review of Thomas's dissent on Techdirt. The dissent is rubbish. Thomas doesn't know what he's talking about. He fails to understand what an API is, for example.
Of course that's not surprising, since Thomas is pretty suspicious of anything that happened after the Constitution was ratified. He certainly doesn't think much of the Bill of Rights.
Breyer's the best of the current panel on IP, I think, and this decision confirms it.
And in many cases they've gone against the expectations of those who nominated and supported them. Souter, for example, was appointed by Bush I, but is generally considered to have voted "liberal" in most of the decisions while he was on the court. (This Wikipedia article, which should of course be taken with a grain of salt, shows Souter as ranging from 54% to 71% liberal in various categories.) Kennedy, appointed by Reagan, was often the swing vote and in his day generally considered the hardest to handicap. Scalia, definitely right-wing (but famously a very skilled thinker, jurist, and writer, and friend of RBG), skewed left on federalism and taxes, according to that same Wikipedia page.
SCOTUS judges do what SCOTUS judges want to do. They're pretty tough to influence, except in the approved manner (i.e. well-founded legal arguments grounded in the legislation and jurisprudence of the US).
The 5100 series wasn't a complete failure. It was the only portable APL machine around (as far as I know), and so had some fans in finance and some other sectors. I knew one or two people back in the day who had worked on 5100-series machines (employer-supplied, of course; in today's money those things would be around $50K - $100K).
If only we had a large corpus of jurisprudence on what constitutes "fair use" in the US... oh, would you look at that.
I think Breyer's opinion is pretty clear. This decision clarifies, not clouds, the status of open-source software in the US. It's not an ideal decision – the court ought to have found APIs not subject to copyright, as Alsup did, and instead left that question open – but it acknowledges the issues with copyright law for software, and draws a pretty clear line regarding fair use.
It's not that odd. The case was an attempt at a cash grab by Oracle, so it was simply a relatively high-risk investment. It didn't pay off, but now that it's over, Oracle's in a more stable position: It won't be spending more on this investment, and there's no possible exposure from setting a potentially damaging precedent. To the extent that anything which happens in the stock market is at all rational, this likely improves the value of Oracle a bit.
It's nice if the instrument lights have a brightness control at all.
Some years back I had a rental Hyundai Santa Fe -- one of those indistinguishable, undistinguished crossovers that have polluted automaker lineups. The lower trim levels of those things are popular with rental companies in the US.
It had the obligatory unnecessary irritainment screen in the middle of the dash, and its backlight did not dim with the instrument panel. So there I was driving through Darkest Kansas in the middle of the night with this brightly glowing rectangle in the corner of my vision. Incredibly annoying.
Of course, it's been many years since I had a rental car which wasn't more or less terrible in one way or another.
Sending a spacecraft to Jupiter and Uranus is much easier than it would be to send a spacecraft to the nearest neutron stars or black holes
Easier, yes. Also, unless you can get a grant with a very long timeline for the final report...
I mean, there might be a neutron star as close as a mere 250 ly away, which is a stroll round the park in interstellar terms, but there's still going to be a bit of a delay before you get information back from your probe. And it looks like the closest known black holes are another order of magnitude in distance from Earth. "Look, when some researcher in some future civilization publishes on this, I want second-author credit."
Oh, ReCaptcha is complete crap. It exists only because Google's found a way to get lazy people who run sites like StackOverflow to get them free labor. It's one of the reasons I quit contributing to SO. (Another is the fact that the site is no longer usable unless you enable scripting, thanks to their use of crap from cookielaw.)
So twenty people here (so far) think machine learning is defined by a labeling procedure with perverse incentives.
I think we can discount the possibility of human intelligence, too. Certainly it's evidence for a paucity in critical thinking.
I think Micro Focus still owns the UNIX copyright. It would have been odd to transfer it to SUSE, since Linux and GNU are not UNIX-branded. (The UNIX trademark and corresponding branding are controlled by The Open Group, of course.) But we don't seem to talk about the UNIX copyright much.
OpenServer 10, a Xinuos invention, is based on FreeBSD; but SCO OpenServer was based on System V, with other bits later incorporated from UnixWare. For this complaint to be meaningful, it has to be about SCO products prior to the formation of Xinuos.
The article says this new complaint cites both OpenServer and UnixWare, so UnixWare is dragged in anyway. UnixWare is the direct heir of SVR4. So we're really talking about an AT&T UNIX heritage here, not BSD.
None of which changes the fact that the whole case is rubbish, of course.
There are many very serious problems with extant facial-recognition and other ML-based computer-vision systems, but what you've written above is a bunch of vacant handwaving, frankly.
Reducing modern ML model architectures to "statistical approximations of numerical digests" is about as meanginful as saying they're "arithmetic". It's not a useful description of the stacked ANN architectures that most of the approaches are deploying, with multiple layers of convolutional and fully-connected networks, often recurrent networks as well, and other functions.
There's no evidence to support the argument that what the human visual system does is qualitatively different from "statistical approximations". And, yes, that includes Penrose's thesis, which is ambitious but fundamentally unpersuasive. (A type-4 doxastic reasoner can believe its own type-4 nature, specifically its inability to believe itself contradiction-free; that's isomorphic to a type-4 doxastic logic system proving its own type-4 nature, and that's isomorphic to proving the Incompleteness Theorem.)
And there will always be a way to fool humans, too.
I'm no fan of facial recognition, but I prefer to dislike it because of real, technical shortcomings, and not faith in hypothetical qualia like "intuitive suspicion".
There are places in the US which still teach COBOL.
Any competent programmer can learn COBOL. The IT labor market is tight, but it's not that tight. And there are plenty of people who could use a better job and would be capable of learning how to maintain a big COBOL back-office application. It's not trivial, but it's not rocket science either.
Typically systems like this aren't just COBOL -- there will likely be some screen definitions (BMS, MFS, whatever), a fair bit of JCL for running reports, quite possibly some 3xx Assembler, maybe some PL/I or REXX, often some COBOL that's generated from a 4GL...
Migrations are often driven primarily by the expense of leasing mainframe hardware and software, and the potential cost of upgrading to newer hardware.
But, yes, rip-and-replace projects frequently fail, and often migrating with minimal changes is a more successful strategy. Of course, the latter pay my wages, so take that with a grain of salt.
Is there though?
Yes, there is. The vast preponderance of evidence and analysis from any number of studies in various disciplines supports that.
Emotional reactions precede logical analysis.1 It doesn't matter, subjectively, that conformance is essentially the same condition whether it follows from the top-down or bottom-up exercise of power; what matters to (nearly all) people is how they perceive that exercise of power in the first place.
1And the former strongly condition the latter, as amply demonstrated by many methodologically-sound psychological studies, and some neurological ones, such as the experiments done by Antonio and Hannah Damasio and their team.
girls in particular will rail against school uniforms and the imposed "uniformity"
I dare say there's ample evidence, including methodologically-sound ethnographic and sociological studies. to show this is not, in fact, a behavior particular to any demographic.
And yet, when they go out as groups, it's not uncommon to a group of girls wearing their own self-imposed "uniform", ie all wearing very similar clothing styles.
There are academic disciplines which had studied this phenomenon at some length, such as Cultural Studies, particularly in its original UK incarnation.1 Or cultural anthropology (Douglas's famous Purity and Danger being one prominent example). Or Critical Theory in the mode of the Frankfort School. Or ethnography (which is sometimes a branch of anthropology, sometimes a type of inquiry or set of research protocols under some other discipline) – I note a few ethnographic studies of teenagers' clothing choices.
Of course, teens are typically under a complex set of strong social pressures, and labor under the combination of a certain amount of personal freedom on the one hand, but numerous constraints that adults are usually free of on the other. And they have immature frontal lobes, which reduce their ability to moderate their emotional responses (relative to neurotypical adults, though certainly many of the latter suck at it too). So their decisions may not always be entirely rational.
And, of course, behavioral economics have shown that people most often assign large weights to immediate and intangible factors when making decisions anyway. Most decisions do not take long-term consequences, and certainly not logical consistency, into account.
In any case, culturally speaking, teenager behavior is something of a self-fulfilling prophecy. The category of teenager is a recent invention – only about a century old – and popular culture assigns all sorts of confrontational attitudes and behaviors to it. Teenagers are routinely urged by mainstream culture to "act out".
1Cultural Studies in the US became rather more diffuse and often less rigorous, and shows less of a focus on identifying and investigating particular subcultural groups and their relationship to mainstream culture. See Pfister, "The Americanization of Cultural Studies".
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
