lynx-dev Licensing Lynx: Summary

[Brett Glass]

Well, I and my fellow developers have reviewed the online responses 
regarding the possibility of licensing the Lynx code for the past several 
days, and have come to some conclusions, which are summarized in this 
message. (I'll draw from several postings to this list in this response.)

As those who have followed this thread will recall, I and a group of 
programmer acquaintances, having watched the problems a blind friend had 
with the Web, seek to produce a Web browser that's truly friendly to the 
visually impaired. Since it's unrealistic to expect that such a product 
would earn much money, we expect to earn minimum wage, or perhaps less, for 
our efforts --- but do need this much income to provide necessary support 
services such as telephone help. (Clearly, a blind user can't turn to the 
Internet for help if the browser is not working.)

Just as others have looked toward published, freely available source code 
such as the Berkeley TCP/IP stack -- which effectively enabled the Internet 
-- to avoid reinventing the wheel, we sought to turn to Lynx for some of 
the code we needed.

Some of those involved with Lynx development welcomed our efforts, but a 
few vociferous ones have decried them. These people seem to have adopted 
Richard Stallman's extremely negative view that allowing one's work to be 
used by someone who derives any form of income from it -- however slight or 
necessary -- it is somehow "selling out." (This despite the fact that we 
weren't asking for "something for nothing" but rather offering to 
compensate authors for the use of their code if they so desired.)

It appears that, sadly, Stallman has managed to convince some credulous 
souls -- via questionable rhetoric -- that it's evil to be paid for one's 
work if that work happens to be selling software. As David Wooley put it in 
one posting:

>If people do decide to sell out on this one (and I suspect not enough
>will) remember that the value of the code to him is probably comparable to
>the labour costs in recreating it together with overheads (payroll tax,
>office space, development hardware).  Although he could develop it himself
>for a similar cost, he would not get to market as quickly and the risks
>would be higher than starting from largely written and tested code.  What
>you would be selling is not the code but a reduction in risk and a quicker
>time to market; and of coarse the ability to profit from his proprietory
>additions.  (You would be no more selling the code than Red Hat sell
>Linux or Microsoft sell Windows 98.)

In other words, it's useful for programmers to stand on one anothers' 
shoulders instead of -- as Brian Reid once quipped -- one another's feet. 
Unfortunately, Stallman's GPL actually prevents good work from taking place 
by locking programmers away from useful code -- creating an "artificial 
scarcity" (something which Stallman, ironically, decries). In this case, in 
particular, it goes even farther: it impairs the development of a useful 
tool for a group of people (the blind and visually impaired) who have 
expressed a real need for it on this list. As Janina Sajka points out:

>The problems you list are not just problems with Lynx. They show up in all
>kinds of applications that blind users have used for many years. This is
>why the various screen readers have developed strategies to handle them
>properly. My Lynx, for example, does not read the status line unless I ask
>for it. It also does not speak the name of the link I'm leaving as I arrow
>down.

Karen Lewellen notes that there are problems with RealAudio (which won't 
work from DOS) and with some screen readers which rely on the PC BIOS:

>I am a blind user and have used lynx either as a part of delphi, or more
>recently as a stand alone driver for a few years.  I know i am only just
>begining to tap into its features, from the stand alone aspect, but the only
>problem i have had, is not really a problem, just a wish.  many programs
>write to the bios, allowing speech to be automatic.  I would love lynx to do
>this, saving me the time in using full screen commands.
>on a different note, can lynx deal with real audio?

This is another common problem, and one of many we'd love to do something 
about. However, it does require the licensing of code from Real Networks, 
and we would (naturally) not have the right to reveal the source for the 
licensed code. Thus, instead of producing a positive outcome for anyone 
(most especially the blind community), adherents of Stallman's agenda throw 
a stumbling block on the way. So eager are they to condemn such efforts, 
despite our good intentions, that we see assertions such as the following, 
from Kim DeVaughn:

>Seems like just an opportunistic "code grab" in an attempt to get some
>sort "enhanced", "blind-friendly" product to market with an FCS (First
>Customer Ship) date right around the time the Fed's new rules are to go
>into effect.
>
>Developing such an app from scratch to meet such a ship-date, is obviously
>not a feasible undertaking ... hence the attempt to co-opt the lynx source.
>
>But perhaps I'm just being overly cynical ... but I doubt it ...

In fact, the Federal guidelines refer to the construction of Web *pages*, 
not web *browsers*. Little work is being done to make truly accessible 
browsers for the blind. However, such cynicism is poisonous to efforts to 
remedy the problem.

A few people on the list have even gone as far as to take pride in their 
desire to sabotage our efforts. Philip Webb writes (to Richard Stallman) 
that "you will be very proud of us all [for attempting to block efforts to 
produce this product]".

The GPL, alas, is designed as a weapon to accomplish just this end. As 
David Wooley writes:

>Commercial prices are set by what the market will bear.  He has already more
>or less admitted that Lynx is in a monopoly position.

This is another unfortunate consequence of the GPL: It is likely to be 
better, in fact, than Microsoft at creating entrenched monopolies. (GCC, 
which is in the process of monopolizing the compiler market even though 
there are superior commercial products, is an excellent example of a 
product which is not as good as what is produced commercially but 
undermines these productive and creative endeavors.) Like Microsoft, it 
gives away products to destroy programmers' livelihoods. All of us -- 
whether users or professional developers -- lose when this happens. We 
become, in fact, more "stuck" than many users currently are with Windows.

Ironically, Richard Stallman refuses to acknowledge that the GPL uses 
Microsoft's own tactics. He writes:

>If a company develops proprietary software, then whether it be small,
>large, or in between, it is not contributing to our community.  If
>they compete with Microsoft by behaving like Microsoft, that doesn't
>make free software any better.

If Richard were truly to take his notions regarding comment to heart, he 
would not advocate the use of the GPL to undermine creative efforts to 
develop commercial software -- open source or otherwise.

Which is what the GPL does. The GPL, in fact, violates the Open Source 
Definition by discriminating against a particular field of endeavor -- in 
this case, professional software development.

As for the legal status of Lynx: At least some of the source files that 
comprise Lynx (though, thankfully, not all of it) appears to be 
legitimately licensed under the GPL, though the package as a whole cannot 
be. (Bellcore and CERN/W3C would be within their rights to protest the 
release of the code under the GPL if they so desired, but only they -- not 
a third party -- would have a cause of action.) Nonetheless, the GPL 
accomplishes its "poisoning the well" effect in that the addition of GPLed 
code muddies the legal status of the whole and hides authorship. (This, 
too, is inimical to the interests of programmers; since they're not being 
paid for code they allow to be freely reused, they should at least be given 
credit.)

Under the circumstances, however, it appears that a few approaches to the 
problem are still on solid legal ground. We can use the Bellcore and CERN 
code without impediments, and are still soliciting the use of other parts 
of Lynx for which authorship is clearly defined. (If anyone who is reading 
this list would be willing to license substantial modules of code which he 
or she has contributed -- anything other than the screen rendering code -- 
we are still interested.) We may hire programmers to "clean room" some of 
the code whose status is less clear, and will almost certainly have to 
rewrite the rest. Fortunately, DEC/Compaq -- an "evil" distributor of 
"proprietary" software -- has been very generous with information and 
sample code for one particular group of output devices (the DecTalk line of 
speech devices), and this will help us. Creative Labs -- another group 
which Stallman would surely condemn -- has also been helpful.

The blind, as a rule, don't have great financial resources (as a previous 
poster mentioned, some 60-70% cannot even find employment), so none of what 
we're doing is any sort of "get rich quick" scheme. We'll probably lose 
money on whatever we do (if we can make minimum wage for our efforts, or 
enough to hire some full-time telephone people, we'll be happy). And the 
prospects for upfront funding are bleak. (A few groups have said they MIGHT 
buy the finished product if they get some funds for it.) So, the GPL and 
the hostile, spiteful attitude behind it (originally spawned by resentment 
of commercial spinoffs of taxpayer-funded academic work at MIT) are 
particularly damaging in this case.

In short, Richard, you've done a great deal to impede an effort to benefit 
a group of people who are in real need.

I hope you're proud of yourself.

--Brett Glass