[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Auth]SEE daemon
From: |
Norbert Bollow |
Subject: |
[Auth]SEE daemon |
Date: |
Sat, 15 Dec 2001 00:43:08 +0100 |
David,
since there is a high level of renewed interest in moving the
SEE daemon forward, could you please post an introduction from
your perspective into this package... what has been implemented
already and what are the next steps?
BTW I'm including (below) a key message from our discussion of SEE
with Richard, which I believe has never been posted publicly so far.
Greetings, Norbert.
--snip------------------------------------------------------------
Date: Wed, 17 Oct 2001 11:51:00 +0200
Message-Id: <address@hidden>
From: Norbert Bollow <address@hidden>
Prefer-Language: de, en, fr
To: address@hidden
CC: address@hidden, address@hidden, address@hidden
In-reply-to: <address@hidden> (message from
Richard Stallman on Tue, 16 Oct 2001 18:43:15 -0600 (MDT))
Subject: Re: A statement on portability in DotGNU
References: <address@hidden> <address@hidden
> > Let us be specific. The only part of DotGNU that really must have a
> > Windows counterpart is DotGNU SEE.
>
> Does having a "Windows counterpart" for DotGNU SEE mean making DotGNU
> SEE run on Windows (as well as on GNU and other systems)? Or is it
> something different from that?
The easiest way to get what what we need will indeed be to
implement dotgnu-see in a reasonably portable way, so that we
won't need too much help from Windows programmers to make it
work well on Windows. But it'll still not be trivial, because
it needs to be "end user friendly".
There are three important areas to consider here:
a) In order for the "virtual identities" system to be useful
for website owners, the client-side parts of it must work
with the vast majority of desktop computers. The current
plans of the various project proposals for the "virtual
identities" system all rely on executing code on the client
computer. This must happen either in the browser (which is
unreliable and a bad long-term strategy as long as Microsoft
has a lot of influence on what capabilities the browser
provides, or doesn't provide) or in DotGNU SEE, our "Secure
Execution Environment"). Execution of code in DotGNU SEE
will either be directly as a plugin (for programs that are
distributed with DotGNU, and which have been compiled to
native code with gcc), or via a Java bytecode or IL JITer
which in turn is a SEE plugin.
b) There will also various kinds of webservices that will rely
on being able to execute code on the cient side. Again,
this code would be executed in the "DotGNU Secure Execution
Environmant". So DotGNU will not be useful for providing
this type of webservices unless DotGNU SEE is made available
for all widespread desktop systems.
c) As we're moving into the area of webservices, it is very
important to keep in mind that the end users should have the
freedom to keep all data on their own computers. This
fundamental freedom is very strongly attacked by Microsoft's
.NET, and I believe that it is very important that we very
clearly address this issue. Microsoft is building a
web-services infrastructure that they control, and they will
allow some programs to be run *only* on machines operated by
Microsoft and Microsoft's "partners". This means that the
customer is forced to transfer to those computers the data
that those programs will process. I believe that this
violates a very fundamental, privacy-related freedom. It
will also lead to many customers not even having a copy of
all their business-critical data, which results in a very
strong case of vendor lock-in. There can be no doubt that
what Microsoft is doing here is outright evil. How will
businesses be able to switch to Free Software if Microsoft
is in control of their business-critical data?
In contrast, DotGNU is designed to prevent this kind of
vendor lock-in, by making sure that all the code which
powers a webservice can also be executed locally, on the end
user's computer, in the "DotGNU Secure Execution Environmant".
This gives the users of the webservice the essential freedom
of keeping their data on their own computer.
The programs which implement webservices under DotGNU will
all be available as Java bytecode or IL on the server side.
Whenever a client is properly authenticated so that it has
the right to cause the execution of the program on the
server, the client can also cause the transfer of the
program (as Java bytecode or IL) to the client-side "DotGNU
Secure Execution Environment".
I feel that it is very important that we can say that DotGNU
protects the freedom of to keep your data on your own
computer. It's a sad situation that currently many people
are not granted the freedom to do their work with a Free
operating system. We can criticise their employers for
that, but at the same time we must admit the reality of this
situation, and avoid a situation where a widespread adoption
of DotGNU-based webservices might lead to these people
losing the freedom to keep the data (which is essential for
their work) on computers that are under their control.
To summarize, those parts of DotGNU which are of interest only
for _providers_of_webservices_ only need to work on GNU systems.
For example, a program for balancing the load between multiple
webservice servers, or the C# compiler, such programs only need
to work on GNU systems. We don't need to worry about porting
them elsewhere.
DotGNU SEE on the other hand needs to be available on the vast
majority of desktop systems, in order for DotGNU-based
webservices to be useful and ethical.
Greetings, Norbert.
--
A member of FreeDevelopers and the DotGNU Steering Committee: dotgnu.org
Norbert Bollow, Weidlistr.18, CH-8624 Gruet (near Zurich, Switzerland)
Tel +41 1 972 20 59 Fax +41 1 972 20 69 http://thinkcoach.com
Your own domain with all your Mailman lists: $15/month http://cisto.com
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Auth]SEE daemon,
Norbert Bollow <=