John Cherry's Weblog

Desktop Linux - still about freedom

Mon, 10 Dec 2007 08:11:48 GMT

Hi freedom lovers everywhere.

The 2007 Desktop Linux Survey results are out and freedom still tops the list for those that are deploying Linux desktops. You have to do a little digging in the results to uncover this salient fact, but when asked about deploying one or more of the pre-installed Linux desktop/client products, the results were that 56.6% of the recipients believed that pre-installed LInux offerings did not meet their business requirements.

Now that several pre-installed Linux products are available, would your organization be likely to deploy one or more of these pre-installed Linux desktop/client products?

Yes, our organization has been waiting for pre-installed Linux offerings. - 42.4%
No, pre-installed Linux offerings do not meet our business requirements. - 57.6%

These responses definitely came from an "enterprise deployment" perspective. When IT decision makers and system administrators listed the issues with deploying pre-installed Linux offerings, they indicated that freedom trumps convenience. When deploying corporate desktop systems, the administrators want the freedom in defining settings and configuration options. The choice of Linux distribution is import as well. IT organizations do not want to be locked into the Linux distribution vendor that is pre-installed on the hardware they have selected.

On the flip side, 42.4% indicated that their organizations have been waiting for pre-installed Linux offerings and these offerings would meet their business requirements. These desktop Linux deployments take advantage of the integration testing, system management, and support that comes from pre-installed offerings with Linux distribution support. However, even with these desktop products the administrators will tweak configurations settings, define pre-loaded sets of applications, and set security and remote access policies (VPN) before deploying the clients. Freedom is still a huge factor in deploying and maintaining pre-installed Linux clients.

This survey was not directed to consumer users of pre-installed Linux desktop systems. When you move outside of the geek and developer circles, everything just has to work out of the box. Pre-installed Linux systems are very important to the consumer circles. They need to be able to power on the machine and be immediately able to play their mp3s, play their DVDs, browse the internet, use their office productivity tools, set up their printer, manage their photos, and download applications that are interesting to them. Most Linux distributions that are targeting the consumer markets support this level of turnkey operation. The latest one that I have tried with outstanding results is a Ubuntu-based distro called Linux Mint.

So who actually took this survey and were there any other interesting results?

The survey was sent out with a shotgun approach, but a profile of the recipients can be derived pretty closely from the responses. The typical respondent...

came from a small company (1-100 people) - 69.4%
is an IT professional or software developer - 43.3%
is involved with computer software companies or education/research - 35.9%
is from the US or Europe (this is the English version) - 86.5%
has already deployed Linux - 64.1%

All the data from this survey is available and you may draw your own conclusions, but I have taken the liberty to draw a few conclusions myself.

For the enterprise client, it is really a two horse race with Windows and Linux. Of the companies represented, 57% were running greater than 50% Windows clients and 46.6% were running greater than 50% Linux clients.
There is extremely high confidence in using Linux for mission-critical applications (76%).
66.1% of the companies responded that Linux was used for client desktops. Many of these are configured as thin clients.
While Adobe Photoshop came in as the top Windows application that should be ported to Linux (47.5%), the majority (61.8%) indicated that their "best" plan was to use equivalent Linux applications where possible.
Eclipse wins (32.7%) as the top developer environment, although it was interesting to see that Microsoft Visual Studio pulled 9.9% indicating that many applications are developed on Windows and ported to Linux. The strength of Eclipse is that it can be used on Windows or Linux.
The top "potential issues" in migrating to Linux is for the support of new devices. In order, missing device driver support was #1, followed by the quality of peripheral support, application support, and the ability to sync with mobile devices.

For enterprise Linux deployments, all the critical applications are available on Linux platforms with the exception of some internally developed applications and some applications specific to business needs. The critical applications were no surprise and turned out to be:

Email - 62%
Browsers/plugins - 48.6%
Office productivity tools - 46.8%
Applications specific to your business - 40.6%
Database applications - 35.9%
Internally developed applications - 30.1%
Secure remote access - 30.0%

For the Linux kernel community and Linux desktop community, the recipients requested that the following items be top priority:

Open source drivers for proprietary hardware - 62.7%
Wireless - 47.2%
Linux desktop standards (cross-distro) - 42.5%

The areas that have gone down significantly in the past two years in their "criticality" include printing (which used to always be at the top), audio/multimedia, fonts and document fidelity, and application packaging. The Linux Desktop community has done some incredible work in these areas.

I won't use this entry to chronicle all the advancements in desktop distributions, pre-installed desktop offerings, application availability, application distribution capabilities, and mobile desktop advances this year. I will talk about these in another posting. However, the data from this survey clearly indicates that there is surging interest in desktop Linux. This was demonstrated by the huge response to the survey.

The survey results are freely available to anyone that wants to use or analyze them. Feel free to feed your closet urges to be an analyst of these survey results. We are all interested in being enlightened by differing opinions and perspectives.

Long live freedom!

Linux Desktop Architects meet again...at Google

Tue, 19 Jun 2007 17:16:37 GMT

DAM-4 Summary Points
--------------------

- The Desktop architects meeting yielded a new set of priorities
for the community. The priorities continue to evolve and shift.
The new priorities set by the Desktop Architects for and with
the community are as follows:

                Packaging
                Multimedia
                Developer Tools
                Power Management
                World Typography
                Printing

- The gstreamer and helix communities agreed to engage in the
   analysis of their respective multimedia frameworks to advance
   the development of Linux sound solutions. Linux distribution
   representatives agreed to evaluate the low level audio interfaces
   and to recommend standard audio and multimedia stack
   implementations. OSS announced that they have moved their code
   into the open source, providing more options in the audio stack.

Breakout notes:
http://www.linux-foundation.org/en/Dam-4_breakout_multimedia

- The LF DAM-4 packaging workgroup included representatives from
the distributions, ISVs, enterprise I/T, individual users and
open source software application owners/maintainers.   The
discussion started to identify the viewpoints of the stakeholders
and what is required to improve Linux application packaging for
all communities. Particular focus on the user experience, how to
manage trusted repositories, package standardization, integration
and testing.   The communities will continue to refine these
packaging requirements on the LSB packaging mailing list.

Breakout notes:
http://www.linux-foundation.org/en/Dam-4_breakout_packaging

- D-BUS was actively discussed by the desktop architects. The meeting
included representatives from both the KDE and GNOME communities
who agreed to issue the following joint statement on D-BUS "Both
KDE and GNOME committed to D-BUS at DAM-4, and to a common set of
interfaces for desktop services."

- The desktop architects agreed to extend standard internationalized
text layouts mechanisms and font management. Adopt HarfBuzz as an
open standard to compliment FreeType and Font Config. (HarfBuzz is
an OpenType Layout engine.)

- The discussion on developer tools surfaced Eclipse as an IDE
solution for the gaps found in other existing developer tools now
widely used by the community, like gdb.

- Recent improvements in power management have shown up to 20% better
power utilization in typical Linux desktops and laptops. Developers
continue to use power monitoring tools to evaluate power usage and
to develop multiple-state power decisions. Good, auto-magic power
management should be pushed into the kernel and into the driver
level when possible, but power policy can be managed in user space.
The general power management case is pretty well covered. However,
more device/peripheral support is needed to take advantage of low
power modes beyond just the power modes of the processor. Power
management developers are meeting next week (June 25-26) in Ottawa
to continue vital power management discussions.

https://lists.linux-foundation.org/pipermail/linux-pm/2007-May/012759.html

- The printing team discussion : Distro independent printer driver
DDK (driver development kit) The DDK will allow printer
manufacturers to target multiple Linux distributions with their
drivers, reducing the time and expense it takes for them to support
Linux. Linux users no longer need to worry about compatibility
between their printer and distribution.

- A record number of desktop architects from the Linux distributions,
hardware and software vendors that support Linux, and open desktop
organizations attended and participated in the Desktop Architects
Meeting, a regular meeting of the open desktop architects.

It was great to see many of you at the Google facility in Mountain View.
It was difficult to catch all the topics, so please add your summaries
on this mailing lists. I'll continue to add presentations to the DAM-4
site as they come in (there were a couple of late-breaking
presentations).

http://www.linux-foundation.org/en/DAM-4_Presentations

Power management discussion will continue next week at OLS. I believe
the sessions on June 25-26 will be led by Len Brown.

https://lists.linux-foundation.org/pipermail/linux-pm/2007-May/012759.html

Relating to world typography, Ed Trager is hosting a text layout summit
at aKademy in a couple of weeks as well.

http://akademy2007.kde.org/codingmarathon/textlayoutsummit.php

Thanks again to the Linux Foundation Desktop Linux workgroup for
sponsoring and coordinating DAM-4. It was a good time for collaboration
for meeting the architects and developers in the desktop community. I
notice many other collaboration meetings outside of the planned
breakouts. When the right people get together, good things just happen.
There were some hacking sessions in mobile space and some genuine
interest from the kernel developers in desktop activities, including
power management. Thanks to the LSB developers that joined our sessions
and participated in the breakouts.

Perhaps the most interesting points relating to the acceleration of the
use of Linux came from our mobile Linux representatives. David "Lefty"
Schlesinger gave a talk on the top issues facing the deployment of Linux
on mobile devices. He basically stated that there were no roadblocks!
Just watch the deployment of handheld mobile devices running Linux this
year. The forecast for the next few years is great. While things like
power management could always be better, there are no real showstoppers
for mobile deployments.

Linux and Solaris - Good luck Ian

Mon, 19 Mar 2007 18:32:11 GMT

With the merger between OSDL and FSG only a few weeks old, our CTO, Ian Murdock is heading to Sun. Ian describes what he will be doing in his weblog...

"And, so, I’m excited to announce that, as of today, I’m joining Sun to head up operating system platform strategy. I’m not saying much about what I’ll be doing yet, but you can probably guess from my background and earlier writings that I’ll be advocating that Solaris needs to close the usability gap with Linux to be competitive; that while as I believe Solaris needs to change in some ways, I also believe deeply in the importance of backward compatibility; and that even with Solaris front and center, I’m pretty strongly of the opinion that Linux needs to play a clearer role in the platform strategy."

I was really looking forward to working with Ian as a colleague at The Linux Foundation, but Ian will remain focused on Linux as a platform and will remain the chairman of the LSB team. If anyone can figure out how Solaris and Linux will play together in the future, Ian can do it. Good luck Ian!

Mobile Linux Platform Guidelines

Fri, 16 Mar 2007 23:18:16 GMT

The Mobile Linux workgroup at the Linux Foundation has as its mission to accelerate adoption of Linux on next-generation mobile handsets and other converged voice/data portable devices. The Mobile Linux workgroup does not impose unfunded mandates upon the Linux industry. Rather, it is creating guideline specifications based on existing open source (prototype) implementations and then works with Linux Foundation members and open source community members to invest in existing projects and instigate creation of new projects to fill gaps and fulfill the promise of Linux in key segments like mobile telephony.

Three of these guideline documents have just been posted

These documents include the Mobile Linux Platform Security Guidelines, the Mobile Linux Platform File System Guidelines, and the Mobile Linux Platform Developer Tools Guidelines. These documents are set up on the wiki as a collaboration point to refine the guidelines and to fill in the gaps. The mobile community is invited to review these documents, collaborate on the content, and to provide feedback to the Linux Foundation Mobile Linux team.

Other documents in the pipeline include mobile guidelines for virtualization and performance.

OSS and the waterfall approach

Fri, 16 Mar 2007 22:25:16 GMT

Project development methodologies fall on a continuum from the structured waterfall approach to extreme programming. I have worked on all of these in my career. The industry is not abandoning all structured methodologies for extreme programming methodologies. Methodologies are being chosen on a variety of criteria. Let me say that the type of development methodology chosen is based on business constraints, span of control, and sharable technology.

If a company is developing a proprietary product that is linked to a business deliverable, the methodology chosen is likely to be a very structured, possibly a waterfall, approach. The process is likely to look something like this...

Company decides to produce product x by timeframe y
Company uses inbound marketing resources to define what the product must look like.
Company puts a development methodology in place, from functional specification to final test that meets the timeframe needed.
Company assigns a project manager to kickoff the process and make sure all the resources are in place.
Company starts the functional, design, code, test waterfall methodology, complete with spec reviews, code reviews, and test audits (would make the ISO crowd happy).
Project schedules start to break down (plethora of reasons).
Process compromises are made.
Product features and capabilities are compromised.
Product quality is compromised...or the product is released as a "beta" for a longer period of time.
Product revisions, product support and product redesigns take much more time and resources than planned. In fact, if the product was not release prematurely with compromised quality, development resources would be available for the next product rather than in fixing the old one.

I left out the very last step because it is very unpopular. Usually the project manager and other innocent developers are casualties when the product blows up in a demo at the trade show. :)

This is a pessimistic view of the proprietary, waterfall approach, but companies continue to use it, they continue to be profitable, and they continue to justify it from a business perspective.

The open source business model says that rather than to have 10 companies developing the same baseline software, let's pool resources and share the commodity components (below the value line). While this pools resources and shares the development load, it reduces the span of control for any one company. Heaven forbid, delivery schedules could be at risk! :) While companies may feel a loss of control, if the components are truly shareable and if the developers have a business or personal interest in the product, the results can be faster development time, better testing, fewer premature deliveries, and more freedom in what can be done to the code (impacts spin-off projects, sustaining, etc.).

Of course, extreme programming removes all constraints and just lets programmers produce code, recode, re-evaluate, recode, etc.. without any real or artificial boundaries. Smart companies that employ traditional, structured methodologies leave some room/budget for open source and/or extreme programming methodologies.

Pre-installed Linux

Wed, 14 Mar 2007 17:41:34 GMT

For a recent major publication, I was asked about Linux being pre-installed on desktop offerings. Mark Shuttleworth, and a host of others in the industry are also responding to inquiries about what this would mean to the desktop ecosystem and what hurdles would need to be crossed by the vendors to make pre-installed Linux a reality. It is definitely not a drop-kick proposition.

When asked about Linux as a pre-installed option for HP, I replied:

HP is considering an offering of a Linux-installed PC as part of its standard offering for one simple reason: Big deals!

HP has been hesitant to jump into the Linux desktop space because in both the enterprise and consumer space, there has not been the demand for it. When we hear that HP is responding to deals involving thousands of desktops, we can assume that these desktops will be deployed in government or enterprise environments. If HP goes to all the work to integrate Linux into desktop hardware, and to provide support, training, and system management for the enterprise, why not create a general offering as well?

By the way, HP has been poised moving into the enterprise desktop market for years. HP is a model community member when it comes to supporting a full line of printers on Linux. An example is the HPLIP support for HP multifunction printers (printing, scanning, faxing).

What factors would make this the right time for HP and Dell to make Linux come standard on a PC?

It has always been difficult to predict which factors and which threshholds needed to be crossed for Linux to be offered as a pre-installed option for desktop/laptop PCs. Linux has a long way to go if we are talking about replicating the entire Window desktop ecosystem. But for many client purposes, Linux is already there. With the Linux platform becoming more standardized (LSB), document formats becoming standardized (ODF), available media formats, and basic office capabilities becoming available as open source (Open Office) and web-based services (Google apps), the Linux desktop is good enough for a whole range of enterprise deployments. AND, while Linux desktops may not replicate the Windows desktop ecosystem, many consumers are seeing that a Linux desktop is good enough for them as well based on responses in Dell's suggestion box.

If large vendors such as HP and Dell come on board with pre-installed Linux would that alleviate user fears?

Support from large vendors such as HP, Dell, and Lenovo would definitely alleviate consumer fear. Consumers don't care what operating system they are running as long as they can do what they want to do with the computer (applications). If you are an average consumer that wants to browse the internet, read email, handle documents/spreadsheets, and manage media (music, videos, pictures, etc.), a pre-installed Linux desktop may be for you. As a bonus, your PC is likely to less expensive and far less vulnerable to viruses and worms.

What are some challenges to getting Linux-on-PC to become common and accepted by the general population?

As I mentioned, Linux is not likely to be a Windows replacement in the near future. It is estimated that over 30,000 applications are available to consumers for Windows platforms. However, as more and more applications become available for Linux, more markets will be open for Linux-based desktops. We have already seen significant penetration into the fixed use markets (point of sale, ATMs, airline kiosks, etc.) as well as technical workstations (software development, animation, movie production, CAD/CAM, etc.) and transactional environments (banking, call centers, travel agents, etc.). Now, we are hearing the starting gun for basic office use in the enterprise as well as some basic consumer markets. Challenges that will be addressed by the hardware vendors delivering pre-installed Linux include hardware support (especially for non-open hardware and the latest plugglable widgets), support for open and proprietary media formats (should be easy in pre-installed systems), and the ability for 3rd party software vendors to easily create and install software on the platform. By the way, the Linux desktop ecosystem already offers hundreds of quality open source applications that exceed the capabilities of much of the proprietary offerings and are certainly more satisfying than the pre-installed "cripple-ware" that is bundled with many of the Windows platforms.