From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00,LOTS_OF_MONEY autolearn=ham autolearn_force=no version=3.4.4 X-Google-Thread: 103376,7684e927a2475d0 X-Google-Attributes: gid103376,public X-Google-Language: ENGLISH,ASCII-7-bit Newsgroups: comp.lang.ada Subject: Re: can one build commercial applications with latest gnat and other licenses related questions... References: <449d2a28$0$11075$9b4e6d93@newsread4.arcor-online.net> <449d5c03$0$11074$9b4e6d93@newsread4.arcor-online.net> <6sbqsh6jv7.fsf@hod.lan.m-e-leypold.de> From: M E Leypold Date: 26 Jun 2006 14:08:14 +0200 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii User-Agent: Some cool user agent (SCUG) NNTP-Posting-Host: 88.72.218.245 X-Trace: news.arcor-ip.de 1151323328 88.72.218.245 (26 Jun 2006 14:02:08 +0200) X-Complaints-To: abuse@arcor-ip.de Path: g2news2.google.com!news4.google.com!border1.nntp.dca.giganews.com!nntp.giganews.com!newsfeed00.sul.t-online.de!t-online.de!newsfeed.freenet.de!newsfeed.kamp.net!newsfeed.kamp.net!news.unit0.net!newsfeed.arcor-ip.de!news.arcor-ip.de!not-for-mail Xref: g2news2.google.com comp.lang.ada:5032 Date: 2006-06-26T14:08:14+02:00 List-Id: Dennis Lee Bieber writes: > On 25 Jun 2006 22:49:48 +0200, M E Leypold > declaimed > the following in comp.lang.ada: > > > > > Completely right. But: There is also the case when the customer wants > > the developer(s) to sign an NDA. Usually (happened too me) these > > customers are deeply uneasy with the GPL around. They fear (and I > Don't know how cost effective it would be -- but, from my viewpoint, It wouldn't. > at least, if a customer wants an NDA in the contract negotiations, the > developer could bring in the cost of the GnatPro license (with compiler > system with license treated as part of the deliverables, or obtained by > the customer with developer working onsite) Well. The calculation would be (as I have shown in another post) that the customer would want to maintain the software for something from 3 to 10 years which with your method would equate to $45.000 to $150.000. Obviously that only has now impact to the customers view on cost effectiveness of the project if that amount is negligible only if the project volume is much large than this amount of money, which would certainly be > $1.000.000. Let's face it: We have a restriction here: Go GPL or pay. I'd wish that people would stop inventing ingenious methods of either circumventing this restrictiction (distributing mangled source or tying the customer into other dependencies that make him refraining from distributing further) or methods of hiding the costs (let the customer pay the compiler support). There is a restriction. It stands to reason that it kills a number of projects that would otherwise been economically viable (to the satisfaction of developer and customer) and can now not be done ore with costs that make them non viable. That much, I think, is undisputed. What is disputed, though, is, wether it should be like this. I cannot say anything on that topic anymore: There are 2 different schools here. One holds that a free tool allows to produce works with the tool that can be sold, even as "closed works". The other holds that "free can only begat free" and that using free tools anyone should not expect to be able to do more than to produce free tools and probably free services. Both schools found their respective view on different evaluations of the situation and probably on different underlying ethical concepts as well (there is always a strong ethical undercurrent in the arguments there). I don't thing it will be possible to mediate between these two schools of thought. Personally I think that "free can only begat free" is contradictory: ACT (i.e.) bases their services on a free compiler, their services are not free though. Same applies to closed source businesses using Apache, Linux, MySQL for working tools. Nobody seriously suggest, that only non profit organisation should be allowed the use of GPL software from now on. All that as nothing to do with copyright, admittedly, but it makes understandable why it is necessary to draw a line between "producing a derived work" (copyright applies, GPL can be enforced) and "using a tool" (that the tool incorporates parts of itself in the product and thus makes it a derived work would be an unwanted effect in this case, hence the birth of LGPL). All in all I don't think that the GPL is being well served by not drawing that line clearly: Much of the FUD around the "viral nature of GPL", like "use a GPL tool and your source must be free" is based on exactly these fringe cases where either the authors of some software didn't care enough to ensure clarify the case or did purposely use the GPL to lock their software away from what they considered unfair profiting. But back to the main theme: I stipulate that the projects killed by the restriction are actually those whose killing harm the case of proliferating Ada most: They take place in the area, where people shift from purely hobbyist or educational use to the first commercial applications of Ada and start to build their infrastructure for future projects. Let me also add: Having tought software engineering courses at german universities I wouldn't try to introduce Ada as a teaching language there today: Whereas the language has all it should have, the skill set acquired by the students would be only usable in very restricted scenarios and this in a small market anyway. It simply wouldn't be responsible to restrict the student's future to the following choices: - Work for DASA or ESA or for the defense industry. - Try to build your own pure Ada shob (w/o previous experience: hard) with whopping fixed costs for the Ada part of your shop. - Restrict yourself to producing GPLed software only in your shops "primary system language". > The GPL curse applies to small independent developers seeking > to produce generally marketable products -- the next Quicken, or > VersaCheck, for example... Given the prohibitive cost of obtaining a > closed-distribution capability for GNAT, they might easily chose to > substitute some other language -- M$ C++ compiler without the > VisualStudio environment appears to be available for the download; > Sun's Java seems to be behind a number of applications... That sums it up nicely. You might substitute "they might easily chose" by "they chose", because that is what the are regularly doing. How do you think that programming languages are chosen? I suggest the following considerations: - Above all: How expensive is an environment with "closed distribution capability" (CDC)? - Can I use the language as single system language, i.e. even if my current application isn't graphical in nature, would I be able to program graphical applications? Would I be able to iterface to the Operating system? - Which libraries (with CDC) are available? - Can I find developers / programmers / collaborators in that language? - How many compilers / build systems are there in the market? Would I be able to fall back to another vendors build system in some years? Would I be able to fall back to community supported software for at least the last years of low intensity maintenance? - And what are the language advantages / disadvantages? The last point creates a kind of vicious circle actually: No CDC -> Skill set not attractive for students -> no Programmers -> no use of the language in small shops --> skill set not attractive for students -> ... You get the picture. Regards -- Markus