Richard Stallman is responsible for the shrinking economy

hi

this current pirate bay thing being case in point. Having 80+ CDs stolen is not the problem, the problem is not getting the 80+ bit patterns back without paying again for them. The music industry/media industry purposefully refrains from maintaining a central repository in order to make an extra profit out of the plastic and paper crap shipped with the bit image.

The bit image is obtainable, so in that sense it was not theived, and pirate bay provides the service of reuniting me with my goods for which I have paid.

Copyright crapmonsters always want to make a profit out of crime, I would suggest a large percentage of their profits are made from witholding the bit patterns of paid up customers.

Even many UK insurance companies place no value on dowloads. Giving much support to the idea that the bit image is valueless. And any enforced fake ideas that they are stealable are grossly for profit out of crime reasons.

cheers jacko

It was confusing for us too ;-)

Ok. If anyone wants to get back in touch (business-wise) with "the cygnus crew" and has problems, feel free to contact me directly.

That makes very good sense to me. I suspect that an employer in the US, who continued to pay an employee to just keep them out of the market place, would find a high degree of legal sympathy were it ever to be the case that such an employee tried to directly compete with the company paying the compensation. So if an employer were to follow what you describe as German practice, I think it would work out for them similarly well.

It's just that a lot of employers believe they have a right for far more than that. And many companies make it a condition of employment to sign such non-compete clauses. In many states in the US they are supported by law in that practice. For varying times, I understand.

I wish it were the case that the US uniformly supported the German perspective.

I can easily imagine a corporation here in the US arguing that this compensation was already provided as part of the compensation in the form of "prior employment." Or, arguing that the act of employment itself was the only necessary compensation. Etc. Without the law being very specific on the terms of "compensation," companies here would argue every possible circumstance they could imagine to see if it "stuck" with the local court system.

Well, that's part of the "boilerplate" (standardized text) of many employment contracts here, as well. And it is broadly enforced, though I can't say how well. My guess is that some trade secrets are more difficult to demonstrate than others, so probably a lot of this is a matter of interpretation.

That also makes a great deal of sense. And restrictions like that have indeed been attempted in the US. I don't think they have succeeded very well, though. At least, I'm not at all worried about those kinds of attempts.

Jon

Hmm, I haven't seen this. Though I usually only deal with compilers for 32 bit machines. The releases tended to be infrequent, and major releases usually had some small incompatibilities. Such as run time library changes, deprecated features finally vanishing, and playing catch up with a moving C/C++ standard.

It also takes an amount of time for a bug fix to wind its way through the vendor's development process and make it into a release that you can get.

I refuse to sign any IP sign-over rights in a contract. I useually put a pen through it. This also goes for any other clause in a contract that exceeds data protection needs and standards.

Sometimes this leads to the other party getting uperty and demanding payment due to cancelation of contract. This is a big problem when I already signed the contract 3 years earlier, and they wanted me to extend it and accept a change in terms.

No competiotion clauses are also a no-no. Share options are the best method of securing such non competion. Me thinks they want there cake and eat it.

Most employers think slavery via contract is a valid method of attracting 'talent' and they are sadly mistaken. case in point is the crunch caused by enforced contractual obligation to make shareholder profit as evidenced in all plc incorperation documents. Sad that such 'profitable' behaviour is higly unstable in some industries/markets, and no non share holder interests (i.e. the public interest) are explicit.

Work my be a means to money for life, but is often a means of money from life. And when the pavement scrape up has to happen, the tax payer is meant to foot the bill. All money comes from somewhere off somebody. Well we all know what it means to off sombody. Like the orders did not MEAN anything, like people are so suprised, like WTF?

Along time ago in a galaxy far far away, on the planet of idiotron, lived a race called the sadomasokistas they queued each day for the privelage of accepting pies for work on the tools of torture of the other que of the contracted invisible ones. Then suddenly the lights came on as the workers demanded more light to do the job, to get the bonus for more productive tool use. They saw the wonder when they discovered that yes it was the others of themselves having the pound of flesh removed, and all along they thought that it was some species of the dark taking pot shots at them, cos it needed to feed, just natural they thought, I mean who'd let a world fall into such ordered repair???

cheers jacko

That's a bit out of date. All versions of the base Qt libraries (but not Qt Extended) are currently available under the LGPL, as well as the GPL and commercial versions.

The commercial version still has the restriction that you cannot use it with code "developed with" the GPL or LGPL versions.

I generally do the same. Usually, that is acceptable. Sometimes, it's not to the other party. If not, we agree not to work together. So I'm with you.

One consideration that I've been asked to consider, at times, is the idea of signing a non-compete merely because I'm acting in a significant technical role and they are seeking financing from investors and need it to shop around. They know, the owners, that it means nothing and that it cannot be enforced, but want it anyway because it sells for them.

I haven't accepted that argument, as yet. But I have had the experience of having it handed to me on a silver platter.

Jon

In at least one (US) case, the courts barred a former employee from working for a competitor, in spite of the fact that his contract explicitly excluded any "no-compete" obligations. The rationale was that disclosure of confidential information would be inevitable.

I think I heard about that. A possible counter to that argument is that if the primary value of a person's skills are narrow enough that he or she cannot possibly get a _valuable_ job without competing (or working for the original company), then it amounts to depriving them of their livelihood.

Jon

FWIW, my preference is for something which gets extensively tested in the real world. Which basically means gcc and MSVC. Nothing else even gets close to the market share of those two.

If your code depends upon standards-compliant floating-point, you shouldn't be using any version of Borland C under any licence.

No argument. I have had my say. Including misspelling 'atrocious'.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

No sweat. It was just a thought.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

... snip ...

I gather you only post on c.a.e, and never read others posts. :-)

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

... snip ....

Omit the "you will". I am far too old and doddering to start such a project. However I am willing to comment and suggest.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

You are aware that using IEEE floating point is an option in the C standard. It is not required. Certain floating point characteristics are required.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

I am aware of that. I'm also aware that just about every other compiler in existence automatically gives you whatever degree of IEEE conformance the FPU happens to provide.

In Borland's case, they just couldn't resist e.g. unconditionally "optimising" x/y to x*(1/y). So while every other x86 compiler calculates

12.0/3 = 4.0 (exactly), Borland gives you 3.99999... This indicates that either they don't understand FP (at all), or at least assume that none of their users might actually care.

gcc doesn't do this at *any* optimisation level; you have to explicitly specify -funsafe-math-optimizations or -ffast-math.

gcc on x86, on the other hand, silently turns 'double' variables into 'long double' when registerizing them, causing code like a = /* some formula */; b = a; /* some code not modifying a or b */ assert(a == b); to fail. This is a flaw of the x87 FPU which must be explicitly worked around using '-ffloat-store'.

But I suspect everyone to have some skeletons in the closet for floating point. For example, older versions of Analog Devices' VisualDSP++ implicitly treated all 'double's as 'float', arguing "you'll usually not need double precision, and it would be very slow because our chip has to simulate it in software and has no 64-bit multiply". Newer versions use the more usual 64-bit 'double's, but as far as I remember, you still have to explicitly tell it you want IEEE-standard floating point that behaves right with all denormal, infinite etc. values.

These are the reasons why I try to avoid floating-point where I can. Most of the time, it'll even become faster this way.

Stefan

GCC is not the only compiler which does this; the Microsoft compiler does this also when compiling for x87 unless the "ANSI" or "Improve float consistency" options are enabled. The enabling one of those options produces significantly slower floating point code. Fortunately these days the x87 is becoming a thing of the past.