It isn't like somebody threw out a specialized acronym and expected people to understand the subject matter instantly. You could get a pretty good guess what was going on from the context and the details are secondary to the story.
After that, if you want to know specifically what Kelvin's Predictor was or how it works, it's not to hard to look it up.
Maybe you paid $12K back in the early Pentium days for a non-Enterprise machine, but the Fortune corporation I worked for didn't. We had that thing hooked up to a mainframe tape drive, which itself was something like $15K in Y2K dollars. We were loading massive quantities of data onto high-speed SCSI disks, and in fact the ONLY reason the whole app wasn't mainframe-based was that the department head was a control freak who didn't want to be at the mercy of the central data center.
For that much money, it was a big disappointment that one of their major selling points never panned out and that the system arrived missing critical functionality. Which, according to the invoice, when I went back and checked - we'd paid for.
Plus, as I said, this isn't the only case where Dell didn't offer the flexibility that I expect from a server-grade machine.
Their "enterprise" level support is excellent and the products generally perform as advertised, if not better.
Hmm. We once paid 12 grand for Pentium server that under-performed the 486-based DFI machine it replaced. Turns out they had forgotten to install the L2 cache chips.
The "upgradeability" of their systems never seemed to pan out. It tended to be based on special form-factor riser cards which rarely, if ever got upgraded models.
I've got a on server sitting right now that I'd love to re-purpose, but the motherboard isn't standard, so putting in a new motherboard and power supply won't let me keep using the rather nice box with the swing-out service gates.
In fact, the only "Dell" I have that's not uselessly unique is actually built on an Intel motherboard, was apparently part of a large lot sold to the US Government (and then retired). and has no documentation at all on the Dell website. Not even a "birth certificate".
That's not doubt, however. That's - as you say - pointless pedantry.
I believe in precision as much as anyone, but they need to tow the line. If we give free reign to endless repetition's of the pointless, their burying us in irrelevancies and the discussion looses it's meaning.
The problem is, most software projects - be they in-house or open-source, tend to have a few scraps of documentation written at the beginning, then as the project takes shape, the basic premises mutate, but no one has the resources to go back and correct/enhance the original documentation. The actual app is full of unstated assumptions because the developers have been doing it so long that they've forgotten that some of that stuff isn't as intuitive as it seems to them. Or worse yet, stuff that's flat out misleading or wrong because of the early-stage changes.
That's a far bigger problem than the relatively smaller changes that come between software releases.
Good documentation requires good coding practices. It's now common practice for the developers to put in JavaDoc/Oxygen/whatever comments into the source code, but the next step up is to remember to explain in the comments what the module is actually supposed to do and give some samples of how its methods are supposed to be called. If you use the JavaDoc system to its maximum potential, you can produce quite literate docs, not just a brute-force enumeration of functions. Few do, however and that includes Sun/Oracle.
These source-bound docs can be used to create the internal documentation set/developer's guide. The next step is to produce usable user documentation. For that, it's usually better to have someone who's a documentation specialist with direct and immediate access to the developers. AND developers who aren't useless jerks who just blow the documenters off with "Read the Code!" AND, incidentally, it wouldn't hurt for code reviews to include comment reviews and witholding integration until the docs are accurate. There's too much artificial urgency in software development anyway.
These days we have really good resources for creating quality docs on a wiki-style platform, so that's the next level I'd shoot for.
Finally, however, a "frozen" copy in the form of an actual book-structured document is important to me. Sometimes I get lost wandering around the hyperlinks on a wiki. In fact, sometimes I'd really actually have a dead-tree reference than even an epub or PDF. Sure it's out of date the minute it's frozen, but the idea is that it's easier to take something that's chronologically consistent and interpolate the latest mods than puzzle through a bunch of stuff that's all over time and space.
2) To reduce by a significant amount. Modern usage.
The Romans haven't been decimating for about 1700 years now. We don't think there's much doubt as to which definition most people are using these days.
why someone can be nice with someone else in person and brutal with the same person on a open mailing list ? It inconsistent and probably a sign of immaturity.
The same reason that perfectly nice people can be real jerks once they get in an automobile.
Best guess is that being immune to direct physical assault brings out people's darker sides.
But programming is simple! A child can do it! Why, just the other day my little nephew Marvin showed me a program that makes a box move around on the screen why going "boop! boop! boop!". Then little Cindy showed me the website she made full of pony pictures! With kittens!
Have you forgotten? We were told that we were now an "Information Economy" and that the blue-collar jobs would all be replaced with new technical sector jobs so nobody would even miss them.
I know, I know. The Roman process of decimation meant "to reduce by 1 tenth". But the modern-day meaning of the word is almost invariably "to reduce TO 1 tenth".
You can whine all you like, but the English language is mutable. "Nice" flip-flops back and forth from being an insult to being a compliment. Political Correctness has done much the same to make word "special" an insult. It happens. Get over it. For the purposes of this discussion there's no ambiguity. Now where's the popcorn?
Do you have any comprehension of the amount of energy stored in a tank of gasoline?
Do you have any comprehension of the amount of energy stored in a 3-oz bottle of distilled water?. E=mc^2. If I didn't slip a decimal, the blast radius should run to approximately 6 miles.
Tab-card machines are digital computers. They operate based on the presence or absence of holes in the cards.
The simplest of the tab-card machines was the sorter. You "programmed" it by turning a screw that moved a wire-brush sensor to one of the 80 column positions, then ran a deck of cards through it. There were 12 pockets and when a hole was detected at the corresponding column/row, a deflector would send the card into the corresponding pocket.
IBM also made more sophisticated unit-record devices, including some with primitive arithmetic-logic units. Those devices were programmed using wires jacked into plugboards that were imprinted with markings that indicated which holes wired what data source or destination. All digital signals again.
Settle down and let the drugs wear off. You're dripping foam on the carpet.
If there's any "whipped-up" indignation I have, it's because the whole Federal process is so infamously turgid that it doesn't support democratic participation and shoulders aside the Main Street businessman in favor of specialized Big Businesses. Basically, if you want to deal with Uncle Sam, you have to dedicate a significant amount of resources specifically to dealing with Uncle Sam. Not to mention a fair amount of expertise in Federal-specific processes and interactions. A pet Congresscritter or 2 won't hurt either.
In IT, depending on where you get your stats, how bad things have to come short to be defined as "fail", and how you define "big", somewhere between two-thirds and 90% of all IT projects fail. That's irrespective of whether the customer is the New York Times, the State of California or the Federal Government. So the idea that simply being a Big Business contracting to the Feds magically makes for success carries no weight with me. The Big Business vendors who have repeatedly failed at some of the previously-mentioned project sites carry Big names like IBM, Oracle and SAP.
There are valid reasons for an entity with lots of experience working for the government to win contracts like this. "Bob's corner firewall shop, viruses killed ded" isn't going to win any 1B multi department government contracts & would fail miserably if they tried.
From what I've seen, the bulk of the "scientific-technological elite" are working 80-100 hour weeks in fear of losing their soon-to-be-offshored jobs and have no time to meddle with the running of the country.
Unless you want to count people like Mark Zuckerberg as "technological elite".
Slashdot Mirror
The irony is that many of the products under development are expected to be used in every corner of the world.
But somehow they can only be developed by bringing people all together in one place.
????
Or you could, you know, look it up: https://en.wikipedia.org/wiki/...
It isn't like somebody threw out a specialized acronym and expected people to understand the subject matter instantly. You could get a pretty good guess what was going on from the context and the details are secondary to the story.
After that, if you want to know specifically what Kelvin's Predictor was or how it works, it's not to hard to look it up.
For those like me who don't automatically know what some random acronym means.
https://en.wikipedia.org/wiki/...
Also, shame on submitter / editor for not including acronym expansion.
Damn. I thought it was the cows again!
Maybe you paid $12K back in the early Pentium days for a non-Enterprise machine, but the Fortune corporation I worked for didn't. We had that thing hooked up to a mainframe tape drive, which itself was something like $15K in Y2K dollars. We were loading massive quantities of data onto high-speed SCSI disks, and in fact the ONLY reason the whole app wasn't mainframe-based was that the department head was a control freak who didn't want to be at the mercy of the central data center.
For that much money, it was a big disappointment that one of their major selling points never panned out and that the system arrived missing critical functionality. Which, according to the invoice, when I went back and checked - we'd paid for.
Plus, as I said, this isn't the only case where Dell didn't offer the flexibility that I expect from a server-grade machine.
Their "enterprise" level support is excellent and the products generally perform as advertised, if not better.
Hmm. We once paid 12 grand for Pentium server that under-performed the 486-based DFI machine it replaced. Turns out they had forgotten to install the L2 cache chips.
The "upgradeability" of their systems never seemed to pan out. It tended to be based on special form-factor riser cards which rarely, if ever got upgraded models.
I've got a on server sitting right now that I'd love to re-purpose, but the motherboard isn't standard, so putting in a new motherboard and power supply won't let me keep using the rather nice box with the swing-out service gates.
In fact, the only "Dell" I have that's not uselessly unique is actually built on an Intel motherboard, was apparently part of a large lot sold to the US Government (and then retired). and has no documentation at all on the Dell website. Not even a "birth certificate".
That's not doubt, however. That's - as you say - pointless pedantry.
I believe in precision as much as anyone, but they need to tow the line. If we give free reign to endless repetition's of the pointless, their burying us in irrelevancies and the discussion looses it's meaning.
The problem is, most software projects - be they in-house or open-source, tend to have a few scraps of documentation written at the beginning, then as the project takes shape, the basic premises mutate, but no one has the resources to go back and correct/enhance the original documentation. The actual app is full of unstated assumptions because the developers have been doing it so long that they've forgotten that some of that stuff isn't as intuitive as it seems to them. Or worse yet, stuff that's flat out misleading or wrong because of the early-stage changes.
That's a far bigger problem than the relatively smaller changes that come between software releases.
Good documentation requires good coding practices. It's now common practice for the developers to put in JavaDoc/Oxygen/whatever comments into the source code, but the next step up is to remember to explain in the comments what the module is actually supposed to do and give some samples of how its methods are supposed to be called. If you use the JavaDoc system to its maximum potential, you can produce quite literate docs, not just a brute-force enumeration of functions. Few do, however and that includes Sun/Oracle.
These source-bound docs can be used to create the internal documentation set/developer's guide. The next step is to produce usable user documentation. For that, it's usually better to have someone who's a documentation specialist with direct and immediate access to the developers. AND developers who aren't useless jerks who just blow the documenters off with "Read the Code!" AND, incidentally, it wouldn't hurt for code reviews to include comment reviews and witholding integration until the docs are accurate. There's too much artificial urgency in software development anyway.
These days we have really good resources for creating quality docs on a wiki-style platform, so that's the next level I'd shoot for.
Finally, however, a "frozen" copy in the form of an actual book-structured document is important to me. Sometimes I get lost wandering around the hyperlinks on a wiki. In fact, sometimes I'd really actually have a dead-tree reference than even an epub or PDF. Sure it's out of date the minute it's frozen, but the idea is that it's easier to take something that's chronologically consistent and interpolate the latest mods than puzzle through a bunch of stuff that's all over time and space.
Or a White House press release.
We know what "decimated" means:
1) To reduce by 1 tenth, Roman military practice
2) To reduce by a significant amount. Modern usage.
The Romans haven't been decimating for about 1700 years now. We don't think there's much doubt as to which definition most people are using these days.
And it runs so well on all those Linux servers!
only viable scripting language choice
FTFY
I used ReXX, you insensitive clod!
"Who judges merit?
The users."
After all, how else would you explain the explosive popularity of systemd?
why someone can be nice with someone else in person and brutal with the same person on a open mailing list ? It inconsistent and probably a sign of immaturity.
The same reason that perfectly nice people can be real jerks once they get in an automobile.
Best guess is that being immune to direct physical assault brings out people's darker sides.
But programming is simple! A child can do it! Why, just the other day my little nephew Marvin showed me a program that makes a box move around on the screen why going "boop! boop! boop!". Then little Cindy showed me the website she made full of pony pictures! With kittens!
That's assuming that a "storage capacitor" and a traditional capacitor are the same thing. I didn't.
Heck, there are already batteries that are more than just simple batteries. For example, ones with temperature sensors for charge control.
Have you forgotten? We were told that we were now an "Information Economy" and that the blue-collar jobs would all be replaced with new technical sector jobs so nobody would even miss them.
I know, I know. The Roman process of decimation meant "to reduce by 1 tenth". But the modern-day meaning of the word is almost invariably "to reduce TO 1 tenth".
You can whine all you like, but the English language is mutable. "Nice" flip-flops back and forth from being an insult to being a compliment. Political Correctness has done much the same to make word "special" an insult. It happens. Get over it. For the purposes of this discussion there's no ambiguity. Now where's the popcorn?
I'm for the Hastur / Yog-Sothoth ticket.
Politics is about as unspeakable as it gets, anyway.
Capacitors discharge rapidly when shorted, that can make them more hazardous than batteries in certain situations.
Also less forgiving of sloppy loads.
An ideal storage capacitor would have some sort of valve mechanism where you could store fast, but limit the discharge rate.
Do you have any comprehension of the amount of energy stored in a tank of gasoline?
Do you have any comprehension of the amount of energy stored in a 3-oz bottle of distilled water?. E=mc^2. If I didn't slip a decimal, the blast radius should run to approximately 6 miles.
Tab-card machines are digital computers. They operate based on the presence or absence of holes in the cards.
The simplest of the tab-card machines was the sorter. You "programmed" it by turning a screw that moved a wire-brush sensor to one of the 80 column positions, then ran a deck of cards through it. There were 12 pockets and when a hole was detected at the corresponding column/row, a deflector would send the card into the corresponding pocket.
IBM also made more sophisticated unit-record devices, including some with primitive arithmetic-logic units. Those devices were programmed using wires jacked into plugboards that were imprinted with markings that indicated which holes wired what data source or destination. All digital signals again.
And $18-25/hr? How is that supposed to give us our Everyday Low Prices?
Settle down and let the drugs wear off. You're dripping foam on the carpet.
If there's any "whipped-up" indignation I have, it's because the whole Federal process is so infamously turgid that it doesn't support democratic participation and shoulders aside the Main Street businessman in favor of specialized Big Businesses. Basically, if you want to deal with Uncle Sam, you have to dedicate a significant amount of resources specifically to dealing with Uncle Sam. Not to mention a fair amount of expertise in Federal-specific processes and interactions. A pet Congresscritter or 2 won't hurt either.
In IT, depending on where you get your stats, how bad things have to come short to be defined as "fail", and how you define "big", somewhere between two-thirds and 90% of all IT projects fail. That's irrespective of whether the customer is the New York Times, the State of California or the Federal Government. So the idea that simply being a Big Business contracting to the Feds magically makes for success carries no weight with me. The Big Business vendors who have repeatedly failed at some of the previously-mentioned project sites carry Big names like IBM, Oracle and SAP.
There are valid reasons for an entity with lots of experience working for the government to win contracts like this. "Bob's corner firewall shop, viruses killed ded" isn't going to win any 1B multi department government contracts & would fail miserably if they tried.
And this is a good thing?
From what I've seen, the bulk of the "scientific-technological elite" are working 80-100 hour weeks in fear of losing their soon-to-be-offshored jobs and have no time to meddle with the running of the country.
Unless you want to count people like Mark Zuckerberg as "technological elite".