D&C Lug - Home Page
Devon & Cornwall Linux Users' Group

[ Date Index ][ Thread Index ]
[ <= Previous by date / thread ] [ Next by date / thread => ]

Re: [LUG] Closed to Open Source. was: Unix file system folder limits



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Monday 24 Mar 2003 4:38 pm, Bill Wilson wrote:
> from bill@xxxxxxxx
> Everyone who works on our products records time spent and the tasks
> worked on during that period. We have data collected over 25 years so
> the figures I quote are accurate. Being a smaller developer our costs
> are much lower than Navision or SAP.

Developer costs is a euphemism for salaries. 
$no_salaried_developers = $no_development_costs;

> Cutting code is approx 10 - 15% of the cost of producing the product.

How much of that is capital expenditure? Mostly salaries?

> Approx 25% takes place in design and specification this cost has to be

That's not usually capital expenditure either?

> incurred before anyone else could start to work on a large scale system.
> This figure would increase if we were to outsource the other tasks. We

Ah, now that's where I disagree. What occurs to me is that there is another 
fundamental difference here - an open source project is not yours to manage 
in the same way as a closed source - the minutiae of the design specs are 
irrelevant as these can be left to various developers to fix in their own 
way. As long as the new engine code is modular and follows existing 
standards, that shouldn't be a problem. So the design burden is CUT not 
increased by having open source developers. If the costs increased somehow, 
no open source project would get off the ground as the number of developers 
on some projects can be VERY high.

Open Source != outsourcing || contracting_out || franchising

What is more likely with Open Source is demonstrated by Gecko. I've simply 
lost count of the number of browsers now based on what is becoming a quite 
ubiquitous engine. Galeon, Phoenix, ...  just today I read about another one 
in LXF - Epiphany. If only Mobius was that ubiquitous! An Open Source project 
is ripe for adaptation to new environments, new demands and new 
architectures, all within the GPL. Once released into the GPL-wild, it will 
pick up a momentum and direction all it's own and possibly quite at odds to 
how you envision it now. After all, who knows more about the needs of users - 
one individual, one closed team or the entire user base? 

Mozilla wanted a full feature set but other developers disagreed and 
browser-only programs followed. Mozilla didn't have GnuPG but did have a 
plugin design so enigmail was born. What can be plugged in can be unplugged 
too.

> know this because we looked at offshore development and ruled it out on
> overhead costs of management. I would think that there would be a bigger

But you won't be managing it (only at an abstract level maybe). It won't be 
YOUR program anymore, it'll be the EVERYONE'S program.

> increase if we were to go open source.
>
> The balance is in testing and documentation all of which would need
> completion before we could ask for support.

Isn't that the whole point of Open Source? That the testing and documentation 
burdens are REMOVED from the original team by utilising the massive number of 
willing volunteers who can test, debug, improve and document individual 
modules or the entire project without any charges to the original team? With 
an Open Source project, you won't be doing the testing OR the documentation. 
So Zero Cost. You'll just have a hand on one or more of the many tillers that 
direct the development and scope of the programs (n.b. plural) that arise. 
There'll be offshoots and adaptations that are simply not under your 
influence at all. If Mobius was GPL, I'd be quite free to adapt it for my own 
needs and package it up for anyone else with similar needs without ever 
consulting ylem. I wouldn't even have to let you know, just retain and 
publish the source code for your original work as part of my own.

I'd always prefer to read documentation written by users - documentation 
written by developers (closed or open source) is universally horrible.

(Which reminds me, I am looking into a Wiki for DCLUG).

Diversification. 
*************
A commercial competitor can try and close down any closed source program 
because there are limited targets - just ask QuarterDeck. (who? see below) An 
Open Source project can have hundreds of siblings and relations all looked 
after by their own independent or loosely co-operative groups of developers. 
Even if some were targeted, loads more copies would still be under active 
development. An analogy:
1 Big fish and 1 little fish in pond. Big fish eats little fish, little fish 
goes extinct. Big fish moves to another pond during the next flood. Repeat.
1 Big fish and 100 similar little fish in (bigger) pond. Big fish eats lots of 
little fish but the remainder survive to breed and continue. Repeat. Big fish 
eventually dies from over-indulgence and goes extinct. (please?)

> Plus Mobius today has no known faults so customers say why pay for
> support if it doesnt go wrong and our experience over the years is that
> customers do not need a support contract. (Ask Rick at Supertramp they
> use it).

It's easy to say if you control the implementations of the software. Allow 
others to download the project and implement on systems perhaps running 
competitor programs or wanting to use it for purposes slightly beyond or 
obtuse to the original purposes and you just may find problems. Problems that 
Open Source developers would solve in a way that a single closed source team 
simply cannot. Mobius cannot have been tested on the sheer number and variety 
of systems that are available for testing any large Open Source project - a 
commercial team simply cannot expect to receive a viable return for setting 
up or monitoring that many diverse yet similar systems. When there is no need 
for a financial return of any kind, that testing becomes available. All you 
have to do is retain the interest of the developer network.

> So we either have to charge for the product or build faults in so that
> customers need support. Given these choices we choose to charge for the
> product and produce good quality software.

Nah. Have you seriously considered a rebuild as Open Source, as I mentioned 
last time? A new base, a new core, new engine. Start again - leaving your 
most valuable existing code to fill the inevitable gap until your existing 
software (let's call it v4) is matched by the new Open Source software (call 
it v6). Every software project on the planet over 1 year old would benefit 
from a complete rebuild. Not a mere port or fork, a darkrooom virgin rebuild.

> Currently to go somway towards open source we are offering Mobius free
> of Ylem licence fees and charge for installation training and support if
> the customer wants support.   I have serious concerns over publishing 6
> million pounds worth of code if its ripped off I cant afford to sue a
> large company.

Then don't. Publish the specifications and let the Open Source community 
design a new, better, faster, version. Or maybe just one with greater 
flexibility and portability built in - just like Mozilla (which is a bigger 
project than some people think - it's a complete suite, not just the browser 
kit - and with a built-in extensibility that is the envy of many other 
projects.)

(Is there ANY evidence that a GPL program HAS been ripped off by a commercial 
operative?)

Open Source is about the freedom of the project as well as the code. If I 
wanted Mobius, let me outline what I'd want.

1. An extensible configuration architecture that can be moulded to my unique 
needs without any bespoke development requests, ala XUL / Majordomo (which is 
written in Perl).
2. Modularity that can strip the program down to absolute essentials so that I 
can run it on a PDA or embedded platform or bloat it out to a full feature 
set suitable for an SGI Altix supercluster, ala CPAN-ish.
3. Choice of static or dynamic library constructs to fit in with unusual 
machines.
4. Full and frank documentation available for all languages written by 
experienced users.
5. Online support mechanisms using ALL free and familiar formats like mailing 
lists, discussion boards, usenet, IRC etc. with responses from experienced 
users as well as developers.
6. A GPL licence - unamended and unabridged.

And, of course, if that last one is met, then I can have everything that I 
want - because if the original project doesn't come up to scratch, a new 
group of developers can take up a new direction.

> Microsoft are now a competitor with Great Plains and Navision so we have
> an uphill struggle with marketing against the big guns.

Linux isn't marketed but is making Microsoft sit up and take notice - even 
receiving the odd alarmed put-down line from high up bods within MS. Linux 
has a strong heritage and a competition-busting price (0.00) that commercial 
operatives find hard to stomach. At present, all manner of statistics are 
being engineered to try to show that 2x0=20 but Excel never was as good as 
1-2-3. 
:-))

Mobius can be relaunched with a similar heritage line, an identical price tag 
and a similar non-plussed response from commercial competitors.


************
For those who don't remember, QuarterDeck was a company selling a memory 
manager for Windows 3 that was miles better than anything Microsoft could 
achieve and they sold copies of their program all over the Windows world 
because of it. Rave reviews, increased performance, increased stability 
(quite something in Windows-land) and in a tiny memory footprint. What 
happened? Microsoft made them an offer they couldn't refuse and took over the 
project - it was never seen again - although Microsoft allege that the 
technology was incorporated into Windows95 memory management. Yeah right - 
plenty of evidence of increased stability and performance in Win95 wasn't 
there. Not.
(IMHO of course.)

- -- 

Neil Williams
=============
http://www.codehelp.co.uk
http://www.dclug.org.uk

http://www.wewantbroadband.co.uk/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE+f1rIiAEJSii8s+MRAqc/AJ9Nadcz0RNKgUW/qk9YEXW5JfT6tQCfQD21
h7xxg3LsdpumcvjJnGxIwZ4=
=wYuV
-----END PGP SIGNATURE-----


--
The Mailing List for the Devon & Cornwall LUG
Mail majordomo@xxxxxxxxxxxx with "unsubscribe list" in the
message body to unsubscribe.


Lynx friendly