Parser and pricing policy and other software issues
Bob LeChevalier
may be in Beta test by the end of the month. Written in C, it is likely
to be available for PC, UNIX, and MacIntosh, perhaps by summer (the
primary qualification for Beta testers is that you have made some
significant efforts at writing Lojban without the parser - we want
people who have learned enough of the language to really hammer on the
parser for the next couple of months, since bugs in the parser and the
grammar are going to be in obscure corners of the language).
I am seeking input on an important question now, though: how much do we
charge for the parser. Recognizing that software is one of the few
areas we can make a little profit to support our activities that don't
generate enough self-supporting revenue (i.e. most of them), we want to
get a good income. But we don't want to price the program so high that
people are reluctant to buy it - the parser is a fantastic tool when you
are learning the language, since you can see right away whether a
sentence or paragraph is grammatical and what is wrong with it if it
isn't.
I also understand that the plethora of U**X machines sometimes requires
source recompilation, so we need to set a price for source independent
of the run-time version. From what I've seen in various price lists,
source tends to rate a premium price, and U**X versions of software tend
to be twice the price of MS-DOS versions. There is also the question of
eventual distribution in the manner of the Free Software Foundation,
though our finances may not allow this at this point. Whether you will
buy the program or not, your opinions on these questions are important,
and will affect the decisions that we'll probably make at LogFest in
June. Note that we are not especially concerned with keeping the parser
source secret, but at this stage we want to keep version control pretty
tight until the final grammar baseline.
We intend to give a discount to people with positive voluntary balances
with LLG, but the amount is undecided. Give the full price you are
willing to pay in the following answers.
So here are the questions. We don't want formal orders, and no
commitments are implied in responding.
1. Will you be interested in buying the parser as soon as it is
available, will you wait for the textbook before buying it, or are you
unlikely to buy it even then?
2. Which of the following would you be interested in, and how much would
you pay for them (a range or maximum price is fine in response)?
a. MS-DOS 8086 run-time source
b. MS-DOS 80286-optimized run-time source
c. MS-DOS 80386-optimized run-time source
d. U**X (specify version) run-time source
3. For U**X, apparently some vendors distinguish in price or
availability between human-readable-source, and encoded-but-compilable-
source which isn't very readable but is only intended to make the
program portable to systems other than the vendor's standard version.
If you specified U**X source in #2, indicate interest and price for
either or both versions. Note that because of version control, we may
be slower to releasing an open- source version.
4. Nora expects to complete the revision of our LogFlash flash card
management software by LogFest. The new version is compatible with a
revised gismu list with much more completely defined place structures
(100 characters vs. 40) that should also be done later this year. Lots
of bells and whistles have been added to support people who have found
our tuned algorithm not suited to their learning style (e.g. if you skip
several days due to illness or other distraction, you tend to miss a lot
of words. The new version allows you to set a maximum word count on
lessons or to skip adding new words to the word ladder. It also allows
you to choose to allow yourself a second try on errors in case you made
a typo, change the number of times you have to practice error words, and
display or print words for review outside of the lesson.
Again, are you interested in this software (MS DOS only for the near
future), and how much are you willing to pay for it? Also indicate
whether you are upgrading from an old (registered or not) version. This
existing, much simpler, version has been priced at $30 for combined
LogFlash 1 and LogFlash 2, which teaches the rafsi. Program size and
recoding effort, and low interest in the rafsi version will mean that it
will be left out of the new package for now. A LogFlash version
teaching cmavo should be available soon after the gismu program, but may
be sold separately. Would you prefer a bundled product at a lower price
than buying the two separately, but higher than the individual program
price?
(Note: the existing LogFlash program will probably continue to be
available via Shareware, but at least for the short term, the enhanced
version will be available on purchase from us only)
5. For MS-DOS users, we are working on translating the original
adventure game 'Collosal Cave' into Lojban, such that it will have
Lojban commands and text (and Lojban-correct magic words), as well as
slightly more powerful syntax processing to support grammatical Lojban
commands. It is intended to be a bit more fun way to learn the
language, though there is a lot of text to translate. The version we
are working on is in Turbo-Pascal, but there are C versions of the
program for the U**X environment that can be similarly modified once we
have the Lojban command structure defined.
Are you interested in this product when it is available (probably this
summer, but it depends on the text translation effort)? Provide machine
and price ranges for this product per the above - in this case, the
MS-DOS version will presumably be available much sooner than any other
version.
Thanx for your input on these questions. You can provide your answers
on prices and machines to me directly or on lojban-list. I encourage
any open debate on this software, our pricing policy, etc. on the list.
We exist to serve the community and this is one of the few ways we can
find out how.
----
lojbab = Bob LeChevalier, President, The Logical Language Group, Inc.
2904 Beau Lane, Fairfax VA 22031-1303 USA
703-385-0273
loj...@snark.thyrsus.com
David Cortesi
>
> John Cowan has reported substantial progress on his Lojban parser, which
> may be in Beta test by the end of the month. Written in C, it is likely
>
> I am seeking input on an important question now, though: how much do we
> are learning the language...
I think if you charge more than copy and shipping costs -- say $15 a disk,
you are cutting your own throat. You WANT people to learn the language.
The language's only hope for continued existence is to develop a community
of speakers. ANYTHING that acts in any way however small to reduce the
motivation of a potential learner has to be avoided. Anything whatever
that makes a person say, "well, maybe next year... I'll wait and see...
maybe after the baby... after christmas... after vacation..."
To me, software under $20 sounds like a bargain; I'll buy it on speculation.
Software over $20 is a purchase I have to justify to myself, to my spouse.
You say: this is our chance to make some money. I say: this is NOT that
chance. There will be NO such chance until you have at least a few
hundred real speakers -- about a decade at present rates of growth!
(To me, the most pitiful public evidence of JCB's self-deception is his
perpetual dreaming that there is a viable market for Loglan products.)
In my opinion you should aim to not LOSE money on distribution (and maybe
a buck a copy to the author just to make him feel appreciated); do that
and you have done very well indeed. And will have met your real objectives,
to increase the spread of the language.
> I also understand that the plethora of U**X machines sometimes requires
> source recompilation, so we need to set a price for source independent
> of the run-time version. From what I've seen in various price lists,
> source tends to rate a premium price, and U**X versions of software tend
> to be twice the price of MS-DOS versions.
You only say that because you aren't on the network! You should know
that there is more FREE, public-domain SOURCE CODE being exchanged over
the Unix network than in all the PC channels combined. I kid you not.
Good software, written in C by gurus, appears free in my email every day.
Several programs A DAY in comp.sources.unix, more in alt.sources, etc.
> There is also the question of
> eventual distribution in the manner of the Free Software Foundation,
> though our finances may not allow this at this point.
You telling me that Stallman and company charge authors? No way; they
are the people who distribute the best, highest-quality C source code
for free. They are the best argument for charging NOTHING for the
Unix source code of the parser.
Understand, I think you can charge (low) for working, stick-in-disk-and-run
software for PCs or Macs. But the Unix source code audience is very
different. They have to put in significant effort to compile on their
machines, and they would typically be running it on "stolen" CPU cycles,
on employer machines. And in my opinion the chance that anybody would
have the knowledge to make a competitive parser and sell it in
competition with you is just astronomically small. For starters, there
is no market to compete for! And second, by keeping your PC/Mac prices
low, you take any possibly motivation away.
As to your questions: I don't think anybody will pay a premium for
80286 code, since the 8086 version will run, and I can't imagine there
being a significant performance difference in parsing single sentences.
I will pay $19 for practically any disk of software for a Mac.
> 5. For MS-DOS users, we are working on translating the original
> adventure game 'Collosal Cave' into Lojban, such that it will have
> Lojban commands and text (and Lojban-correct magic words), as well as
> slightly more powerful syntax processing to support grammatical Lojban
> commands. It is intended to be a bit more fun way to learn the
> language, though there is a lot of text to translate. The version we
> are working on is in Turbo-Pascal, but there are C versions of the
> program for the U**X environment that can be similarly modified once we
> have the Lojban command structure defined.
This should be a very enjoyable, helpful learning tool. $19 a disk
and it's a barn-burner. But jeez I wish you guys would start using C...
Turbo's variant Pascal is just totally unportable to anywhere...
/////// / David Cortesi ////// cor...@informix.com //////
////// // // //
//// / /// Informix Software // Tough times never last, //
/// / //// 4100 Bohannon Drive // but tough people do. //
// / ///// Menlo Park, CA 94025 // -- Joe Morgan //
/ //////// (415) 926-6300 //// ...pyramid!infmx!cortesi ////