Re: Ada is getting more popular!

["Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr>]

Le Mon, 25 Oct 2010 13:00:19 +0200, Ludovic Brenta  
<ludovic@ludovic-brenta.org> a écrit:
I may reply to others later, but I wanted to reply without waiting to your  
message.

> I think we'll agree that that's a distorted
> relationship between producer and consumer.  In the "sell software,
> not services" model, the software is a vehicle for advertizing and
> selling the services.
Agree and strongly feels this is distorted too. And to be honest, that is  
true for every thing with some more or less relationship with the web (and  
most software are distributed on the web). This has another effect : the  
value is not the value of what is primarily exposed as valuable (the  
software, or the document on web-site), as this is valuable for the  
author, only via indirect ways which has nothing to deal with what is  
exposed as valuable.

> If the software is proprietary, you cannot tell
> how buggy it is or will be in the future and you must depend on a
> single supplier for the services.  If the software is Free Software,
> there is no catch and no hidden bugs; all the bugs are in the open
That is entirely false. Few people audit sources, and open or close source  
does not matter for 98% of users. It is commonly known that with about  
70/80% of open-source software, you have a single developer and nobody  
else care about the source. Further more, nowadays applications tends to  
be so much complex that understanding or auditing the source requires a  
enormous personal investment, so just a few people do that. And this  
complexity helps hidden bug or even back-door. Sorry, forget the link, but  
I swear I recently read about an open-source application with such a  
back-door. Free does not guarantee anything. Bugs may still be hidden by  
lack of human investment and high complexity.

> and
> you don't have to buy services from the same people that provided the
> first dose.
The trouble here is more with a lack of interoperability. The matter more  
deals with public standards, like POSIX, XML, vs proprietary stuffs, like  
the MS-Doc proprietary format.

> No, the service is not secondary to the product and you know that. Why
> else does your company pay for AdaCore's services if they can get the
> same compiler and libraries for free?  Even if your boss only pays to
> "cover his ass" (i.e. help with certification), that is still a
> valuable service.
AdaCore is company to which you could not compare any other. Their clients  
are big clients with special needs. For most others I would say a software  
is good if it does not requires support, because it is fine as-is without  
support. For these common cases, if you want people to rely on your  
services (because you must give the software for free because you do not  
have the choice due to the context), then you must try to be unavoidable,  
a kind of monopole or a kind of unique quasi-actor in an area. Otherwise,  
if you do read-to-use-without-need-for-more-support-software, you simply  
loose. In some way, this context push complexity and trouble-inside  
software, just because support is the only way to get any earn. And doing  
so, you must be sure nobody could provide the support except you (thus,  
the need for high complexity). That is distorted too.

This is not applicable at all to tiny company or stand-alone authors.

> You still have not provided a satisfactory explanation of why you
> think that the "sell services, no software" is socialism, so the
> comparison with the USSR is meaningless to me.
Well, I agree the word socialism may be a bit miss-used here, but that is  
another story (not for that place).

> No, I explain these things by plain and simple greed and short-
> sightedness from most people managing developers and from most
> developers themselves.  Socialism has nothing to do with that.  Proof:
> the miserable state of software you describe exists also in (some)
> very capitalist companies selling high-end, expensive proprietary
> software, especially those who charge only a small fraction of the
> license fees in "maintenance fees".  These companies are simply not
> interested in fixing the bugs that affect their customers; they are
> only interested in selling "upgrades".  And the unrealistic time-to-
> market pressures (e.g. arbitrarily fixed release schedules) make it
> nearly impossible for them to ship bug-free software in the first
> place.
If they do not care, that is not because they sell free or none-free  
software, that is because they are big and as such, they are unavoidable.  
They know whatever they do, clients will be sticked with them (the case  
with MS, as an example, and I say that although I am running Windows). And  
unfortunately, the context push this kind of big companies, because  
alternatives will not have even a chance to born or to survive. As I  
sometime say, “Linux did not help to get ride of Windows, it instead  
helped to enforce MS monopole”. Why ? Because it helped MS killing any  
potential MS concurrents.

The idea that the author as no right at all and only the consumer as all  
the rights, could not ends into anything else. You must have the power of  
a monopole to get a chance to survive in such a context. Otherwise, no  
right for authors and all rights for consumers, simply means you better be  
a consumer than an author. Result : more and more consumers, less and less  
authors (or less and less authors who survive). The result is poor authors  
and poor products, as Dmitry argue.

Note : AdaCore is a kind of UFO here, as it provides a rather good  
product. But I still believe its special status (like public funding while  
others did not get the same, and GPL) made some collateral damage, and I  
know someones (at least one), seems to have suffered from that.


-- 
Si les chats miaulent et font autant de vocalises bizarres, c’est pas pour  
les chiens.