About Scott C. Lemon

I'm a techno futurist, interested in all aspects of humanity, sociology, community, identity, and technology. While we are all approaching the Singularity, I'm just having fun effecting the outcomes of the future!

Confused about the GPL

I had an interesting conversation last night with some Novell folks
about Open Source and the GPL. They were indicating that iFolder
had been put into “Open Source”, but that they would also be selling an
“Enterprise Version” of the product. Sounds good … a way to
make money on Open Source.

What surprised me was that they then indicated that they were not going
to release the “Enterprise Version” into Open Source. What?
I thought the GPL had the “viral” effect. Nope, they said …
they are retaining the copyrights. On top of that, they are
requiring anyone contributing to the projects to sign over their rights
and copyrights. Whoa … now this sounds weird to me.

I
have to admit that I don’t understand a lot of this, however this seems
extremely counter to the GPL. I’m going to investigate more,
however it appears that the following is somehow the path that Novell
is pursuing:

  • Release projects into Open Source under the GPL
  • Retain all copyrights
  • Require contributors to transfer all rights to Novell
  • Release proprietary, commercial versions of the software for money
  • Include Open Source contributions in proprietary, commercial versions
  • Prevent anyone else from creating commercial versions using the GPL

Is this correct? Can anyone really do this?

They indicated that MySQL, SleepyCat, and others are doing exactly this … I’m going to dig further …

Meeting Miguel and Nat …

As I was leaving the Mono GUI session, I went to say hello to Calvin
and he seemed shocked to see me.  It was actually funny … the
Mono team had been reading my blog in the back of the room and Calvin
told them that he knew me.  They had no idea I was itting in the
class in front of them!

Calvin took me out in the lobby where I got to meet Miguel and Nat, and
took a little time to chat.  Of course, Miguel indicated he was
sorry that I worked for SCO … however we talked about Mono, Simias,
and iFolder.  They are really doing some great work and want to offer a real competitor to Longhorn … before Longhorn.

Novell acquired some real talent with Ximian … these guys are going
to make a real difference in the world.  The most impressive thing
that indicates this to me, is that they don’t waste time talking badly
about Microsoft, etc.  Them seem to really respect with Microsoft
is doing, and what Microsoft is creating … and they are confident
that, even with all of the money and marketing power that Microsoft
has, they can beat them.  They are committed to create something
that sells itself … that provides real value that people will pay
for.  And they seem to realize they don’t have to trash Microsoft
in order to do that.  Good people …

BrainShare 2004: Mono: Rapid GUI Application Development on Linux

I didn’t get a chance to blog this last night … this was the last session that I attended Tuesday.

I spent a little time in the lab since this session was marked as full
… I was on the waiting list.  I ended up at the room about 30
minutes after it started and got in anyhow.  As I walked in I saw Calvin Gaisford
from Novell.  I had worked with Calvin at Novell and he is a great
guy … and extremely talented.  He gets it.  He bailed to
Caldera when things sucked at Novell, and went back a year or so
later.  He’s very much involved with the iFolder project, and is working with Mono.

Just as I got in the room, one of the people at a PC got up and left
… I had sat next to him in the morning … he saw me standing in the
back and offered his PC to me.  Jackpot!

I flipped through the slides, which were a good overview of  the
.NET architecture, and then got caught up to the development
projects.  Again … I’m impressed.  I worked through the
first couple of exercises, and then got to the Glade
development.  Glade is an awesome tool that I found a long time
ago in Linux … it allows you to rapidly build the UI framework that
you later add code to.  Mono has implemented libraries so that you
can now use Glade to create your UI, and then add the C# (or VB?) code
later.  GUI apps in GNOME, on UNIX or Linux, written in C# or
VB.  Amazing …

Ease of application development on any platform, IMHO, is key to
success.  And these guys are moving rapidly to provide a full –
modern – development environment with rich tools for the creation of
these applications.  Mono rocks!

During the session the speaker introduced the “crowd” in the back of
the room … Erik the PM, Miguel the leader, Calvin of iFolder,
etc.  I wanted to get a chance to meet Miguel and chat with him
… when I completed the development exercises I turned around and they
had left.  Bummer.

Anyhow … I can’t wait to continue to explore development with Glade
and Mono … I want to see how quickly the progress with VB
occurs.  IMHO, VB is a *huge* with for UNIX and Linux when Mono
supports it … there are a *lot* of VB developers out there.

The Mono sessions alone made the conference worthwhile!

BrainShare 2004: Business Intelligence with Hyperion and the Novell Secure Enterprise Dashboard

Hyperion seems to have some very powerful tools for Business
Performance Monitoring.  I want to chat with them about our
WebFace solutions for creating powerful new monitoring
applications.  It would really fit well

Dashboards … I agree with the speaker about the power of 
“Business Dashboards”.  He is indicating that Dashboards require
the ability to display information, and also provide controls to effect
change.  He also distinguished between Dashboards and Scorecards,
and also talked about some of the overlap:  Alerting,
Benchmarking, Metrics, Graphics, Reporting, and Metadata.  He
feels that they are both combining into a single solution.

Most of the demonstrations that he is showing are static HTML pages of
data and images.  The rep from Hyperion indicated that they tend
to always go to HTML due to the advantages of “no install”.  He
said they do some level of Java applets, however they are minimal.

I’m going to go to the lab to see their demo … they indicated it’s a little thin …

BrainShare 2004: Introduction to SUSE LINUX Server, the Universal Operating System

It was funny … at the beginning of this session was the first time
that I paid any attention to the first slide … the Novell “one Net”
vision slide.  I hadn’t really heard it yesterday … this time it
was with a German accent so I took notice.

The vision of Suse, is to build one universal operating system for all
platforms.  This is not so far from the original NT vision at
Microsoft … cross-processor OS.  I remember when I was
consulting and we tested and installed NT on Intel, Alpha, MIPS and
other RISC processors … I thought it would be impressive.  In
the end, only the Intel platform survived.  There was no real
market for the more expensive RISC machines.

The presenter touched on UNIX history, and the fact that POSIX
compliance really drove forward levels of interoperability.  SuSE,
he says, is the first to then take this even further with one source
code base across 32/54 bit processors, and all platforms.  
He also indicated that their AutoBuild process builds, certifies, and
tests the various “flavors”.

It’s weird, but this part on the build environment of SuSE is going on
and on.  People are leaving … it’s like the presentation could
have been called “This is how we build SuSE products”.  I’m going
to stay a little longer to see where this goes …. I thought I would
hear about features, etc.

He finally moved on to YaST and the power of the feature set. 
Security was next … network monitoring, file system monitoring,
encrypted file system, ACLs, 128-bit SSL, etc.  High-availability
also.

Scalability was next … CPU-scaling via 64-bit processors and 64-way
SMP.  Storage scaling via Multi-path I/O, 1024 NAS-based file
systems, more disk (up to 2,000 devices for 292TB).  Application
scalaing by upping the total memory supported.

Other things … Logical Volume Manager, Journaling File Systems, more Open Source packages.

What’s coming?

  • Kernel 2.6
  • Multinode failover
  • CGL 2.0 (Priority 1 features)
  • Redundant Scalable paths to storage
    • HD mirroring
    • Cluster file system
  • Cluster volume manager
  • Improved performance of large machines
    • SMP, RAM, I/O
    • Native POSIX Thread Library
    • Restructured I/O Subsystem
  • Up to 128 CPUs per system
  • More efficient large RAM and NUMA system support
  • Hotplug and Persisitent device names (sysfs, udev)

There were also various security additions, adding a CA, new GUI admin interfaces, etc.  They commented on using OpenSLP.

Now he’s flipping through slides like crazy …OpenExchange has a lot
of improvements, but I couldn’t really keep up.  On the Desktop
there is a push for other application compatibility, single sign-on,
and some SAP integration.  More hardware support – biometric
devices, mobile power management, and thin-client support.  Lots
of Desktop lock-down features.

BrainShare 2004: Discovering the Future of the Linux Desktop

I’ve been reading about Nat Friedman for a while, and reading his
blog. He’s a good presenter, and a really smart person. He
is doing the presentation so far, and has been showing examples
of the evolution of the desktop on Linux. Starting in 1992,
he showed just how crude things were … and then moving forward to
1995, 1997, and then to today. He feels that Netscape on Linux
was one of the biggest things to drive the adoption and usability of
Linux.

2001 brought Mozilla, OpenOffice, Evolution, and GNOME/KDE. These
first releases were used and the feedback was absorbed through
2002. 2003 the next real iterations arrived of all of these
projects. 2004 is bringing all sorts of new efforts and
contributions to the desktop.

He commented on one of the largest installations of Linux
desktops … in Spain. 400,000 Linux desktops installed in
Extramadura and Andalucia. He commented on the classic issue of
new installs vs. migrations … new installs are so much easier.
He also pointed to the Google Zeitgeist
page where stats show that Linux is still at ~1% of the machines
querying Google … he wanted to now what the rounding is … does the
1% really mean .8% or 1.4% … what’s the real number? 😉

Nat then showed a Linux Desktop Scorecard where he rated various
aspects with A+, A, A-, B+, etc. The worst aspect (rated a D) is
“Application Availability” … he stated that Linux still doesn’t have
all of the apps to do your job. Not in all cases, but many.
Related to that is the Application Interoperability issues. You
might open a Excel spreadsheet in OpenOffice and it doesn’t work.

They outlined the “Linux Desktop” as being:

  • Groupware – Evolution
  • Web Browser – Mozilla
  • Office Suite – OpenOffice
  • Other Components – iPrint, iFolder, and more

He commented on a number of books … one in particular called “The
design of everyday objects”. He was joking about the “Apply”
button in Windows … and that in the Linux desktop they chose to just
have things occur when you select options. He demonstrated
what appears to be a MVC architecture behind the desktop configuration
settings … showing that when settings are altered in one interface,
they are reflected elsewhere.

He commented on Novell’s Open Source efforts, and the “Unification” of
the GNOME and KDE desktops. They also emphasized their efforts to
contribute to Open Source development based on their own internal pain
… for example contributing to Mozilla to make it more IE compatible.

Beating Longhorn to the punch … they are working on Desktop search
(iFolder/Simias), Rich widget toolkits (Avalon/XAML, Cairo, XUL,
Mozilla, Gtk), and High-level multilanguage runtime (Mono/OSS Java).

Talking with Nat after the presentation, I asked him about kernel
dependencies … he indicated that he could not see where there would
be any kernel specific code introduced.  So (per the GNOME
homepage) it will remain that “GNOME is a Unix and Linux desktop suite
and development platform.”  More UNIX-compatible software …

BrainShare 2004: Mono: Developing and Deploying .NET Applications for Linux

The first session that I attended this morning was a Mono “hands-on”
developer session. I have to admit that I continue to be
impressed by .NET and C# … and the Mono implementation.

In the session we actually looked at and created some basic .ASP
scripts and ran them using the Mono server. Overall … it just
worked. The presenter did explain a lot of what still isn’t quite
there … but the basic C# is working, and the breadth of database
connectors is growing.

I’m going to continue to experiment and follow this project … it is
also a UNIX-compatible project … fully cross-kernel … nothing tied
to the Linux kernel.

BrainShare 2004: Open Source with SOA for Rapid Web Development

This session seemed a little weak … probably because I am too
involved in both XP development processes, and web services..  It
started off with a consulting/product company – Picasso Software
– going through some very high-level “case studies”.  Really they
were just bullets and stories of the successes … some good
high-profile customers.

Now I’m watching the second presenter talk about exteNd Director and
Composer.  These are very similar to work that we are doing with
our Host Encapsulator and WebFace solutions.  I believe that the
“legacy application to web services” aspects are nearly
identical.  He is now showing the UI generation, and it is very
“HTML page” oriented (they even call it PageFlow) and based on
XForms.  He did show (and answer my question) that they are able
to create composite web services in their tool … although they are
not yet BPEL compliant.

The third presenter is now talking about what Novell is
delivering.  He is promoting the Novell Development Center … an
actual facility where customers can (pay?) go and develop a solution
for a real business problem.  They offer business and technical
experts to assist.  This appears to be an aspect of their
consulting services.

BrainShare 2004: Introducing Mono for Developers

I got in the room a few minutes late … the presentation had already
began.  One thing right off the bat … it is truly refreshing to
hear this conversation.  So far the speaker has been very
complimentary to C# and .NET … expressing the power and value of what
they provide.  He expressed the adoption of  .NET and C#
within IT organizations, and the broad availability of information,
training and certifications.  He spent some time to express the
advantages of .NET and C# over C/C++ and also over Java.  I agree
with him on all counts.

In discussing the “Internal Impact on Novell”, he addressed the value
of Mono … extending Novell’s application reach to many new operating
system platforms.  What he expressed is his understanding that
abstracting away the operating system – and the kernel specifically –
is a valuable offering.  Mono is allowing Novell to provide
solutions on both the NetWare kernel (I am going to look for actual
examples of this in the lab!) and the Linux kernel.  In
addition,  it appears that this would be a route to get these same
applications to FreeBSD, Mac OS-X … or other UNIX systems.

He talked about MonoDevelop … the Mono IDE.  It looks nice, however I was confused since Novell seems to have committed to Eclipse … I am hoping they are just going to integrate the functionality into Eclipse.

The Mono project is governed by three different licenses: GPL, LGPL,
X11  Different parts use different licenses.  They expect to
release v1.0 this summer … C# compiler, VM with JIT and pre-compiler,
IL assembler/disassembler, development and security tools, .NET v1.0
and v1.1 APIs, etc.

It was interesting that he took some specific time to address that this
is “*NOT* the Novell commercial product … this is v1.0 of the Open
Source project.”  I am curious why this distinction … so Novell
is going to sell and charge for their Mono product?  It will be
interesting to watch.

I asked about support for Mac OS-X and the speaker indicated that he is
running Mono on Mac OS-X right now!  So it is going to be (and
stay?) a fully cross-operating system solution.  Nothing specific
to the Linux kernel here …

The demo was great … but I’ve already been playing with Mono and am very impressed …

Brainshare 2004: ROI of Open Source on the Desktop

This was an interesting session … the presenters addressed a lot of
issues.  They indicated that it used to be about the costs and
availability of support for alternative desktops like GNU/Linux. 
Now, they say, the real issue is the migration cost.  It will be
interesting to see what they say are the biggest on-going costs …

One thing they mentioned was that you get saving through less
staffing.  This was due to adopting a “Universal OS” …
committing to only running GNU/Linux.  I’m surprised by this …
since it seems that IS&T in most organizations will always end up
running a wide range of systems … many different types of 
operating systems also.  I can get in very small businesses and
even medium businesses where there is some flexibility, however most
will have a variety for some time.  Does this mean that if the
customer chooses to use Windows, Mac, UNIX, and UNIX-compatible
solutions there is not the ROI that they promise?

They made a very good point about legacy documentation compatibility
and support.  They indicated that most organizations do not
require the ability to do “mass conversion” of stored documents, but
can implement a plan to do conversion on demand.  They are
promoting .pdf as a good format to convert to … something that can
not be altered easily.  I would think that most documents are for
archive purposes, and that newer solutions – like Wiki software –
provide a better solution for “living” documents.  They also
pointed to a Microsoft URL
that shows a Microsoft XML document standard.  They indicated that
you have to be careful of “advanced” features and their use … these
(I agree) are where the lock-in occurs.  I was surprised they
didn’t mention the Oasis Open Office XML work on this …

There is an issue with Macros, and the use of VBA, in Office
files.  They said there is no easy way to convert these, and that
they do tend to comprise a significant corporate investment. 
Again, this is an area where there is significant (at times)
intellectual property and business logic that has to be
preserved.  I would tend to agree that you want to extract this
from the Office environment anyhow … or look for ways to standardize
and make the code more portable.

As for Migration Approaches, they indicated to core strategies – “Rip
and Replace”, and “Go Forward Migration”.  With their Go Forward
approach they talked about the issues with vendors and applications,
and suggested that a “dual boot” solution might be

In their case-study – OpenOffice.org – they showed graphs where they
believe they will see a 5-year Net Present Value, 84% Internal Rate of
Return, and a Payback Period of 30 months.  I asked them about
what the biggest real contributor to this is, and they said the
Microsoft licenses.  This seems to be a little light since
Microsoft can always drop prices …

They closed with a pitch on Ngage consulting services … I’m not sure if they included this in their case study costs …