Follow Slashdot stories on Twitter


Forgot your password?
DEAL: For $25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 Internet speed test! ×

Comment Re: Why not? (Score 1) 361

Fragile? When I work with 'older' systems, I tend to see a level of 'robustness' that is missing in the newer stuff.

That said, poor development - regardless of language - leads to exactly what you described.

And OO doesn't overly complicate stuff ... it just seems to simply not work for overly complicated systems :) It has real, solid benefits, allows for faster development, more modular programming ... and, once you get beyond the point where one individual can keep everything straight, seems to devolve into a mess faster than any older language.

I like automation. Don't really want to be coding record lengths and block sizes and alignments in some combination of JCL and COBOL. But ... I also like having the option to dive into that level, when necessary.

I am in the over 50 crowd. That said, I'm using SAS (all-time favourite cross-platform language, with a nod to Java for trying to take that title :) , Unix, Netezza, to put together nice databases for people to play with, mostly using Tableau (the current glossiest of the BI set). Doesn't hurt that I know Essbase ... which officially became Oracle BI, but half the developers ended up at Tableau.

I have a pet theory, that new languages are developed when some n00b starts in one language, curses it for not being able to easily do what they want, and creates their own ... without learning the subtleties of the first.

I remember the first time I hit VB ... which was immediately after working on a large CICS project. Don't have to code every !@#$ing location for every field? Don't have to ensure all your calls (and returns) are valid? Drag and drop development? Code that automatically executes, without the user needing to hit the enter key? Awesome! And then ... the downsides slowly became apparent. Debugging was a nightmare compared to mainframes, as the code was scattered/buried all over the place; performance was !@#$. (Speaking off, decades later, a rather large company I worked for replaced their CICS system with a 'Visual' system ... and were frantically adding servers to get it running. Final analysis showed that the new system - regardless of other benefits - required 50 X more processing power).

Who here has played with Panvalet? Yeah, originally designed before mainframes even had 'PDS's (folders) - so it offered a pseudo-PDS, and VSAM files couldn't be over 4 Gb - which it had a work-around for. Eventually, the underlying technolgy (IBM) incorporated those items. By which point Panvalet had a spectacular version control. Who here is using a *separate* version control tool? Currently playing with Tortoise SVN ... and cursing the lack of some functionality Panvalet had decades back ....

There's this idea that technology can replace the need for skills. Seems to be prevalent in upper management, and in any company pushing their technology. And ... pretty much BS. I've seen (and done) systems designed by two or three people, running on near antiquated hardware, leave the 'official' projects (meaning $100M+ budgets, teams of developers, hardware that you have to show ID before seeing) in the dust.

Currently dealing with a large push to HADOOP. Woo! New buzzwords! How's that line go? "All groups develop their own language of obfuscation, to show who is, and isn't, in their group." Minor details like database design are completely alien to the group putting it together. Normalization? Huh, what's that? Slight problem in that at least two of the data feeds are both huge, and contain massive duplication. One system (M$ SQL Server) is already dying, because the same group didn't bother with any of those pesky considerations ... and are a) running out of space and b) users are a little upset that a million dollars of hardware is giving performance you'd expect from a 286 after a lightning strike. I made the joke about 'The key,...,so help me Codd' to the Hadoop developers ... and got blank looks. When I can do the calculations on a piece of paper, showing their current approach is going to snarf their memory in less than two years, you'd think that would be a wake-up call ...

... but no.

I *don't* know that there is any driving need to stay with COBOL. Apart from - generally - it is easier to modify a well-coded COBOL system, compared to nearly anything else. That said, how many 'well-coded' systems are there?

It's like assembler. If you know what you're doing, one can code stuff to run faster. But, is it really worth it? The extra time spent in dev. The steep learning curve when bringing new people in. The near inability to do any coding in it without near toxic levels of caffeine in your system.

A lot of the complaints - and desire for 'new' technologies - have far, far less to do with the language/technology itself, and more to do with how well - or poorly - the previous system was implemented.

There's another quote I love - There are two kinds of fools: those who say 'We've always done it this way, so that is best', and those who say 'This is new, so it must be best'.

Comment Re:SAS get sued so often it boggles my mind (Score 2) 161

Worked at a number of companies that use SAS ... and none had sued them. Part of the love users have for SAS is the user survey: what the users want determines what SAS works on for the next year.

And ... MBAs? WTF? Seriously, SAS is one of the most comprehensive, technically oriented languages out there. It doesn't support the 'McDonald's Burger-Flipping Developer' approach like many other languages - it requires someone with more than a one-or-two year college course to work effectively in it.

It is also, hands down, the best language for cross-platform work. It doesn't promote the 'lock-the-suckers-in-to-our-proprietary-model', promoted by so many others.

There are some 'weaknesses'. It doesn't do maps well, it is pretty poor at graphical stuff, and it's OLAP attempt, while nice to work in, isn't that powerful ... but that is also where adding another software product comes in. As an example pulling data together from multiple sources/platforms, creating a database, then dropping something like Tableau on top of that. You *could* do something like this with OBI ... but the back-end part would be significantly harder than in SAS

Finally ... backwards compatibility. SAS has not screwed its users, with an upgrade that breaks previous code.

I like higher level languages ... the more automation, the better. SAS continues to improve in that direction. At the same time, I like a language that allows one to do things at a lower level, when necessary. SAS has kept that too ... which puts it waaaay above certain other languages/packages in that regard.

Comment Re:Leave. (Score 1) 433

I had a situation where someone went out of their way to backstab me, from outright lies in the HR 'Peer Review' system, to attacking me in a large meeting for a program of mine that crashed, claiming she had told me to fix it months ago.

Took all of about three minutes to hack the HR system. Discovered other people had complained about this individual - repeatedly - for lying.

The version control system showed the fix I *had* implemented had been removed ... while I was on vacation, by this individual.

End result? After showing all this to my boss ... nothing. I lasted one year at this place. My health suffered. Evidence didn't matter. And the individual is still there, 26 years. Toxic as heck. I've spoken to others who left because of her. Again, emails, evidence ... nothing happens. Don't know why HR goes out of their way to continually dump 'the troublemakers', when they all are having trouble with the same individual.

Not to say documenting things isn't a great idea - it is - but, often, HR simply doesn't want to do anything, because that would be an admission there *was* a problem ... and that alone can come back and be used against them in court.

One addition - put down in email anything that is done over the phone ... 'just to confirm my understanding of our conversation'. I've dealt with a few individuals who use the phone, almost exclusively, as a means to prevent their own accountability.

Comment Re:Encrypt! (Score 1) 394

Two thoughts/questions:

What's to stop people from going full on proxy/Tor ? The only IP addresses recorded are worthless.

Is there any limit to how much information each ISP must record? The articles talk about 'A Year'. Can someone write a small app to simply run through a list of a couple of dozen websites - all with extremely long URLs - for a couple of hours after you finish browsing, racking up a Gig or two of storage? Just a thought for something that would make this technically unfeasible.

Comment Re:could just be the beginning (Score 1) 153

Yeah, I kinda threw that in as a complaint, but you're absolutely right: it doesn't support concurrent processing very well at all. Then again, I've seen one installation where it was used only to pull info during the day, and updated (batch) at night. They simply created a couple dozen copies of the database itself. Which isn't all that different than one of the Oracle options. Although, in this case, the developer put a tricky little bit in, so each DB had one user at a time. Insane, and not particularly scalable, but it worked just fine for this instance.

Comment Re:could just be the beginning (Score 2) 153

From my experience, I'd guess that about 90% of Oracle installations do not need Oracle.

I'll go one step beyond that: in my experience, 99% of Oracle installations could be replaced with SQLite, MySQL, Firebird, even Derby. (Possibly Excel, in some cases)

Virtually every Oracle DB I have encountered has used the POWER of Oracle (TM) as an excuse to skip putting together a decent schema. Massive duplication of data. Joining dozens of tables to get commonly needed data. Tables with far too many fields.

I'm currently dealing with one that works ... just. Minor issues, like virtually every table can be dropped by at least two orders of magnitude in size, the actual Oracle DB supporting the application uses ~140 tables, when it needs ten, and there is a ton of data stored that is inaccessible using keys.

If management would spend a fraction of the amount they spend on hardware on a decent DBA, then they wouldn't need to spend millions on monstrosities (in terms of overkill) like Oracle ... and the hardware needed to run it on. Have similar feelings towards Hadoop ... yay, it's great, sooo scalable! Whadya mean I can't get the data out in a usable form?

I do have a grudging respect for OLAP, in one regard: put together a decent schema, and new elements can generally be added by inserting one row in the description table, not by changing the schema itself. That really lends itself to ease of maintenance. But, again, it does require *some* up-front design work.

Comment Re:If confirmed, does this make it realistic? (Score 2) 477

"The 1.2mN/kW1.2mN/kW performance parameter is over two orders of magnitude higher than other forms of “zero-propellant” propulsion, such as light sails,"

I have a question: What the smurf does this mean? Seriously?

Are we talking about a ground-based laser pushing it? In which case, the idiocy of comparing a system that you have to lift into space, in which every gram is critical, versus something here that can be hooked up to the power grid, is beyond belief.

Are we talking about a solar sail that is simply power by the sun? In which case, given the power source is external, and unending (well, good for the next 5 billion years or so), how do you make a claim that something is 'two orders of magnitude higher' in 'performance parameters' than the solar sail? Calculations?

When the paper is 'proof of concept', but a) they don't actually do any experimentation to see what would change performance, b) have lots of explanations for all the possible sources of error ... but, again, don't actually monkey around with said sources, to see if their hand-wavium is correct (apart from the torsion pendulum), .... and c) finish with a unsupported statement claiming superiority over 'other' zero-propellant system ... honestly, they did some great science is some respects, and utterly abysmal in others.

To be fair, they could easily be using previously published numbers on solar-sail efficiency. And their numbers could all be spot on, not to mention their conclusions. But failing to have a paper proof-read by someone NOT familiar with the subject is bad (and all too common).

And they could have made the paper better, but some decent editing. OK, so you start with a thrust-to-power ratio of 1.2 mN/kW. The error margin, ±6 N, is buried way, way down. Seriously, putting the two together would give some real validation to the idea the thrust they got was far more than the (calculated) error margin.

Comment Re:Really? (Score 2) 432

One thing that is missing from the comments is that the counter to anti-ship missiles - decoys, phalanx defense and anti-missile-missiles - are currently looking pretty good, and with the advent of some defensive lasers, improving a lot. The improvements in the anti-ship missiles is mostly a matter of guidance; the basic tech has ben stagnant for a while. In military terms, something is 'obsolete' for just as long until someone comes up with a counter. Air-to-air missiles were initially touted as ending the dogfight era ... then somebody came up with decoys.

Currently, there is no defense against cannons. So if a ship were equipped with only cannons for offense, and all the defensive toys, it would realistically be able to go tot-to-toe with a conventional ASM armed opponent ... for a fraction of the cost. It sucks, but most ASMs are in the millions to tens-of-millions dollar range, while AASMs are in the tens to hundred thou range ... decoys, chaff are hundreds to a couple of thou ... both the phalanx CIWS and laser defense system (e.g. HELLADS) have a pretty large cost for the weapon itself, but the cost per shot is pretty minimal (a few second burst from a phalanx system is a few hundred dollars; the laser defense costs less than a dollar per shot).

Which is how the pendulum swings. Once the defenses get good enough, someone comes up with a better nut-cracker

Comment Re:BIG (Score 1) 90

I think the MCU deal - read: DISNEY - is a good thing for Netflix. One of the golden rules of Hollywood: Don't fuck with the mouse.

Netflix got strong-armed by Comcast a while back. With Disney on their side, Comcast is going to think twice before pulling such crap (illegal, but no charges laid) again.

On top of which, Disney has a long history of using independent subsidiaries to produce their 'riskier' content, thus keeping the 'Disney' name pristine and family friendly. Netflix is truly 'independent' :)

And, of course, Disney owns ABC. While the old networks are struggling to stay relevant, and slowly crawling into the 21st century, Netflix is there. *Any* partnership gives Disney/ABC an edge over the others. There were rumours that ABC shows would be streamed on Netflix, too, similar to how the network sites release them after they air (usually either a day or week delay). That hasn't happened - yet - but I wouldn't be surprised if it does.

Comment Re:State sponsored corporate spies (Score 4, Interesting) 469

We had an ugly situation locally, where a supremely over-qualified graduate, from a top-tier university, was passed over for even an interview, and sued. Born Chinese. The company in question does sensitive work, and had run an extensive program to detect leaks/spies ... and every person they identified was Chinese. They started running the same process on new hires ... and, over a five year period, every Chinese hire turned out to be a spy. So the company simply stopped hiring Chinese. At some point, you can sympathize with their position: why the eff are they spending huge amounts on this aspect of security, when simply saying 'No Chinese hires' solves most of the problem?

It sucks, but unless the governments start treating corporate espionage seriously, and make the penalties serious enough that people won't engage in this behaviour, it is going to continue.

The other issue is that even second - and sometimes third - generation Chinese are leaned on, because they still have family back in China. Again, really sucks, but companies are just protecting themselves.

The question becomes, at what point does 'Not hiring Chinese' go from discrimination to simply safe practice? There isn't a clear answer :(

Comment Re:What about the NBA? (Score 1) 469

Technically, there are no human 'races', as we can all interbreed. However, I'll forgive your lapse :) That said, there is one ethnic group - identifiable by both genetics and culture - that actually does have a statistically significant higher IQ: the Askenazi Jews. And yes, there have been several studies, so that qualifies as 'reproducible'.

Comment Re: Dey tek er jebs! (Score 4, Insightful) 332


In part, it's all about how things look on the budget sheet. Replacing one North American worker for two Indian workers - and paying less - looks good. And the numbers can be shown to management. The downside - inferior code, taking longer to produce - isn't captured as neatly. And the numbers can't be shown to management anywhere near as easily.

And one other fun fun fun detail ... managers get promoted based on the number of people they manage, not the total salary of their underlings. So replacing your home-grown, competent North American worker with multiple lesser-skilled, lesser-paid foreigners means the managers get bumped up a pay-grade.

So ... while the outsourcing (or, in the case of H1-Bs, in-country outsourcing) means that companies pay much, much more for the same software, the people making the decisions don't care about that - they care about promoting themselves.

And one final candle on the cake: the stock market punishes companies that deviate from the pack. If one company were to stand up and say "Hey, this outsourcing is costing us more! Let's stop doing it!" then their stock would take a hit. And corporations are run by the board, for the board: the largest part of their remuneration is stock options.

Comment Re:Denormalize (Score 1) 674

Gotta say, one interesting benefit of working on a Netezza machine (massively multi-parallel) is that tables can be normalized to BCNF ... and joined (as views) so the users never realize they aren't looking at anything but a nice, fat, logical, (business) understandable table.

Not to mention, gets rid of all those pesky nulls. Netezza, of course, being one of *those* systems that recognizes true/false/null ...

Comment Re:The balance of value.. (Score 1) 128

IIRC, the requirements for a frontline TSA grunt are a high school certificate ... or a few months in a 'related' occupation: read security guard. This means any high-school drop-out, with a few months as a mall guard, or night watchman, is qualified. And, the TSA refused, repeatedly, to fulfil their legal obligations, under FOIA, to disclose what the breakdown (graduates/drop-outs), citing 'SECURITY!!!'.

Pathetic. Then again, un-tested x-ray machines, and you are forbidden to carry any radiation monitoring device. Passengers caught with such things can expect to miss their flight, at a minimum.

A group so under-qualified for what they are supposed to do it is staggering. And, when their failure rate is over 98% on any tests, their solution is "We need to hire more people!".


Slashdot Top Deals

The IBM purchase of ROLM gives new meaning to the term "twisted pair". -- Howard Anderson, "Yankee Group"