[REBOL] Re: Licensing, components
From: atruter:hih:au at: 20-Sep-2002 16:22
Thanks for all the positive feedback. Now some responses to responses (is
this turning into a Q&A or what? ;) ).
> These were good comments and a good read. Thanks for your efforts in
giving us this information. - Paul Tretter
Thankyou. Also thanks to Bohdan "Bo" Lechnowsky for actually answering Jason's questions
first! ;)
> Hi, just one question about this: Do you use a Windows compatible GUI
layout or do you use the Rebol look? Later you wrote that you don't want to
let others know how your app was made. If you use the Rebol look it's clear
that this is not a "native" Windows application. Robert
I use the REBOL look (most notably the btn's) but with my own "dialog
management" interface. Basically a title-bar with "Info", "Help" and
Close
btns right-aligned. Clicking on the title-bar makes the client area
disappear which is handy when dealing with images (it lets you quickly see
the underlying image without dismissing the active dialog).
Anyone familiar with REBOL would recognise the UI for what it is, but that
does not make for too many people at this time so I don't needlessly
volunteer that bit of information. As REBOL and / or my product gain market
share then this becomes less important to hide. Someone, somewhere will
eventually say, "Aha, you wrote that in REBOL and *that* is why your
development cycle is quarter the time of ours.". By that stage I should
have a two-year lead on the market (albeit a small one), and as easy as
REBOL is to learn and use, it takes at least a year (IMHO) to tool-up to
the level where you can write decent-sized applications.
> Of course I'd love to hear more here about your application. Can you
describe it better in general or Rebol terms without revealing too much to
your potential competitors?
Sure. I'm not that worried by my competitors as I have a reasonable
technical (both hardware and software) lead on them and the intangible
advantages (marketing, agency agreements, backing of industry and field
leaders) counts for more in the long run.
In general terms the software is an Image Management system that enables a
practitioner (Optometrists at the moment) to capture high resolution
digital images (~5 megapixel) and assign them to a patient and a
consultation. These images can then be searched, compared, annotated (both
graphically and textually). The system allows images (and their "effects")
to be cross-referenced to each other and reports generated via a
sophisticated pdf-maker.r interface (with due credit and copyright
acknowledged).
It is a pure REBOL app (apart from some aspects of the PDF interface) and
could not have been done without the help / knowledge gained from this
list! This is why I am more than happy to "give back" my fledgling
commercial experiences ... if it helps others commercialise REBOL then we
all win.
> You mention Medical Imaging.. what quantity of image data are your
handling, how big are the files, how fast is it, format?
I should have said "Healthcare Imaging" as "Medical Imaging" means
something else to most folks, ;)
Images are ~5 megapixel JPEG with an image quality of ~95% (about 1.2mb on
disk, 15mb in memory). Each practitioner generates 25-75 images per day.
Most image processing (zoom, grayscale, annotation, next, prev, etc) is
instantaneous given a decent FSB (Celeron / Duron are *slow* for this
reason). Emboss and drawing [on the image] tend to be slower, but still
tolerable.
> Interesting. What kind of physuical package do you sell?
Software - CD plus printed manual in a box.
Hardware add-ons include "Image Processing" PC, Digital Camera, Camera
adaptor, Slit-lamp, Beam-splitter, USB / Remote switcher.
The hardware tends to sell the software and vice versa, but the software is
the main game.
> Does this mean your prudct is also not network-oriented either?
Fully network capable. Includes a concurrent licence manager and a
semaphore locking facility (for DB and log operations). I used to work for
Oracle so this sort of stuff is pretty close to my heart. ;)
> The big ongoing debate here is how to stimulate developer needs and
growth as well as profitable REBOL sales. Any more thoughts on that?
I could have written the app in Delphi (I'm not a C guy, although my PERL
is pretty hot, and Kylix lets me write to Linux if need be). Main reasons
for going with REBOL were:
1) The language allows me to implement a large variety of functionality
without recourse to external libraries, add-ons, modules, API's, etc.
Having the entirety of the solution implemented in REBOL means that I can
focus my expertise in one area (REBOL) and have the ability to control and
rectify any functional / logic problems (ie. I don't have to delve into
other peoples code / binaries).
2) The interpreter / compiler must be stable and any bugs must have
simple work-arounds. If it crashes or leaks memory I'm not interested.
Fortunately, the few minor bugs that REBOL does have are easily
worked-around.
3) It lets me develop applications in about quarter the time I am used
to. With REBOL I often find the line between proto-typing and development
quite fine. As the NIKE saying goes, "just do it".
4) The highly compact and readable code size lets me change and fix code
orders of magnitude faster than I could before. Being able to implement a
major functional change hours after it is suggested tends to astonish most
folks (both IT and non-IT alike).
5) A "support area" with a low noise to signal ratio. This ml is probably
the best I have ever come across.
6) A low entry cost for commercialisation. I actually like the royalty
agreement as my exposure in the case of failure is just $499, while 90% of
any success is pretty reasonable to me if the previous points hold true.
Moral issues aside, I would not even consider 10% on a VB or Access app as
the "value" proposition (subjective I know) is different.
I believe "my market" wants something that:
1) works and works well
2) is simple / intuitive to use (KISS)
3) is smart (eg. autosave, postcode completion, etc)
4) runs out of the box on any PC
5) runs on any flavour of windoze without *any* changes / differences
6) integrates seemlessly with application required devices (eg. a digital
camera)
7) generates professional looking printed matter that can faxed, emailed
directly to clients / peers
8) uses common data-storage standards (eg. JPEG) to ensure ready access
to data over the years
9) is secure
This reflects my market. Your market may be Mac (Academics), Linux (Geeks)
or whatever. The more technical your market the greater the functional
expectations.
> What kind of dalog do you have with RT?
Nothing mysterious here, just a few emails back and forth with Cindy on
royalty / licence / encap issues. Trying to start my own company has made
me realise what "time poor" is all about. There is no time to waste. For
this reason I try to keep my correspondence with RT to a concise minimum.
Neither they or I can afford the time.
RT's attitude to changes in REBOL itself are that the more you bring in the
easier (cheaper) it is to get changes. I have no problem with this, I don't
expect my $499 to get me much input into where REBOL goes ... make a
million though and the relationship changes. This is no different from
Macdonald's listening to its bigger franchises in preference to it's
smaller ones.
> Have you directly suggested improvements like these to them?
Yes, mostly encap related.
>> Purchase encap for each platform you intend to market for. Note that
higher
>> sales will effectively let you obtain encap on other platforms "for
free".
>> eg. $10,000 in sales covers the cost of encap for two platforms, even if
>> most sales occurred on only one platform.
> Sorry I don't quite follow your point..
My fault. Let's say I purchase encap for windoze ($499). I can sell $4,990
worth of software before having to hand over additional licence fees. Say
sales reach $4,000 and look set to triple. Say I also wish to purchase
encap for linux. Well if those two things do happen ($12,000 in sales and
purchase encap for linux) then I have paid $1,200 to RT. Whether this was
one encap for windows licence and $701 dollars in commission or two encap
licences (one windows, one linux) and $202 in commission is a moot point.
In both cases I have handed over 10% of my sales proceeds ($1,200) but in
the 2nd case I have two encap licences for my troubles.
My original point was that multiple OS encap licences don't "cost" anything
(just like the original licence) so long as your software sales are 10
times the net amount you spent on encap licences.
> What do you think about an attractive bundle Encap price? Any 4 platforms
for $ abc?
Good idea as the current scheme encourages you to focus on one platform
only (at least until you can "cover" the cost of another one).
> What about enabling limited trial/demo options for Encap?
Probably not a good idea as there are so many ways to implement this,
ignoring the variable policy requirements (expires after 30 days, or 20
runs, or on 1-Jan-2003, etc) I imagine most people will want this feature
(if used) to be exposed through their own UI with their own processing
branches (online register, email, phone, fax, etc).
>> They love the UI of the application and don't know [or care] that it was
written with REBOL.
> Does anyone else?
Had a competitor at the trade show get a stooge to pose as a sales prospect
and ask, "What was this written in?", to which I responded "C", then,
Which vendor?
.
My answer was mostly truthful (I believe the REBOL interpreter is written
in C, which accounts for 99% of the executable size), but demonstrates that
people want to know how the distinctive "look & feel" were achieved.
Remember that most folks writing windoze apps with windoze tools aren't
graphics artists and find it difficult to create a non-native looking UI.
Your typical Optometrist however much prefers an application with a few
clearly labelled buttons as opposed to another windows like application.
They want to show their patients an impressive looking diagnostic tool not
another piece of PC software, I can't stress how important it is that
people want to buy and use simple tools that work reliably ... no-one is
impressed with complex menu structures that offer a multitude of options
and choices of which only 20% are regularly used. KISS.
> Roughly a year ago I exchanged some e-mails with Cindy, and I got the
sense that many of the details
were still being worked out. - Ed O'Connor
RT are doing *something* with their licence arrangements (I don't know
what) so are probably not pushing encap / royalty agreements as hard as
they could. I needed something ASAP (so an agent could demo our executable)
so I went for the Royalty Agreement instead of waiting until "things get
sorted out". I have believe that RT will act in good faith whatever happens
and if different commercial licence arrangements come out in the future
then I would seek to standardise on them regardless of whether they were
more or less favourable than the present ones. My view is that the current
arrangements are interim to a certain extent (heck, my demo uses VIEW/Beta
when it comes to that, but RT Betas are better than most vendors
Production
releases anyway).
> I think Ashley's post is probably the best source of public info about
Encap now. Which suggests REBOL's apparent lack of uptake isn't entirely
due to the commercial REBOLs not being freely is atributable. - Carl Read
Thanx. I hope RT steps in if I have inadvertently misrepresented anything
or anyone! ;)
Regards,
Ashley
