Killer apps/"selling" points of FC and GNU/Linux

Mike Hearn mike at navi.cx
Tue Nov 16 21:01:26 UTC 2004


On Tue, 16 Nov 2004 16:52:08 -0300, Avi Alkalay wrote:
> Java is an open specification with a closed source JVM implementation.
> Sun actually likes projects like Kaffe, and other opensource JVM
> implementations.
> Same openness for Inte/AMD and IBM PC/Whitebox.

Well, if by open specification you mean documented then yeah but so is
Win32 ;)
 
> What I know is that lawyers advised IBM to not offer Wine solutions. Its
> almost the same with OpenOffice.org (for other legal reasons), but for
> this last one the value you get is soooo big that we are fighting these
> lawyers. I'm not saying that Wine doesn't have that value. I use it
> everyday and it works well. But IMHO, I'd like to not depend on it in
> the near future.

Well, this shows that lawyers are usually very conservative. Why? Because
if they warn people of possibly impending doom and nothing happens, they
don't get fired. If they don't warn people and then doom does happen, they
do. Open source projects are all together in the same boat.

> The so called FUD was not intentional. I apologise. I'm just telling you
> the experiences I've had with customers. Expading my point: companies
> are buying the idea that server-side applications centralization and
> integration, with a web/portal approach, is the right way to go, so a
> Wine-based solution feels like temporary, and sometimes does not worth
> the effort and migration costs.

Ah yes, web apps. This works for some things. However:

a) Not all programs work as web apps. Two examples I've worked on lately:
   medical hardware monitoring software (written for win3.1, no source
   code!), electronics design software (graphics intensive)

b) Convincing companies to rewrite debugged and tested software as web
   apps is pretty tricky

c) No guarantee web apps are portable. Internet Explorer is one of the
   most popular apps to run on Wine, and we (CodeWeavers) routinely
   improve it for corporate customers. Why? Because they have web apps
   that were written using ASP and require IE. Sad, I know .... can't
   make any assumptions about the competence of the people writing the
   web software though :(

> Yes, it is a rocky path. But (real story follows:) imagine a company
> that has 7000 VB apps spread all around. An IT services company can't
> say "Lets test your apps on Wine, and then we'll know if it is good for
> you, and give you a contract to sign for the migration service". It is
> just too risky and too much (free) pre-sales work that nobody can
> afford.

Yep, I know. It is very tricky indeed. However, you don't have to do it as
pre-sales. Look at it this way:

* You don't have to migrate the whole company at once. Desktops can be
  moved department by department. It's *very* unlikely that every user
  needs every app, so you can take it slowly over a period of years.

* It doesn't have to be pre-sales. If the company has 7000 VB apps (I can
  easily believe this is true) and they want to migrate to Linux they are
  either going to have to rewrite them or use Wine. If they work out of
  the box on Wine then great, you only have the cost of testing them. If
  they don't you have the cost of making them work. You would have had
  cost anyway though due to the rewrite. Using VB apps on a non-Windows
  platform will always have cost. This applies even to rolling out service
  packs so IT departments are used to doing these calculations.

> Great. This means that Wine has an excelent ecosystem and it is mature.
> I hope lawyers can change their opinion about Wine, because it is a
> great open source project.

I suspect the only way to change lawyers opinions is to prove them wrong,
judging from past history ...

I agree this is off-topic but at the same time, I feel strongly this is a
conversation the community needs to have. After all there's no point
writing an OS if hardly anybody uses it, we'd be the next BeOS! So I think
I'll hit send anyway and feel guilty for a bit ...

thanks -mike




More information about the fedora-devel-list mailing list