Anthony Barker's Weblog on Linux, Technology & the Economy

eWeek OpenOffice review

Here's my take on eweeks latest Ooo review.

OpenOffice.Org 1.1.1

Cons

# File-format compatibility issues Although OpenOffice.org does a good job of handling Microsoft Office file formats, small formatting inconsistencies will require reworking of complex documents.

Agreed - this is due to proprietary microsoft formats. Its not really a con - Its a problem of Microsoft Lock-in.

# Lack of traditional support Office suites typically do not require much vendor support, but the fact that OpenOffice.org is an open-source project means software support must come from the community, generally spread out across various Web sites and newsgroups.

Disagree - There are several vendors offering support. For purchased support Sun as well as Redhat and Suse.

# Interface differences OpenOffice.org is similar to Microsoft Office in its design, but users will need some time to grow accustomed to differences between the two.

Disagree - this again isn't a downside of the product, but rather a product of Microsoft Lock-in.

My cons:
* Ooo's gui is not pretty enough.
* Would like to see integrated addressbook/email
* Wordperfect + Lotus 123 filters

For a list of the bugs in 2.0

Office 2003

Pros

# Familiarity Most knowledge workers use some version of Microsoft Office already.

Disagree - this again isn't a downside of the product, but rather a product of Microsoft Lock-in.

# File-format compatibility Microsoft Office file formats are de facto standards.

Disagree - this again isn't a downside of the product, but rather a due to Microsoft Lock-in.

# Advanced features Office 2003 has more features and capabilities than competing suites. Although many users do not require or use much of this functionality, advanced users, particularly of spreadsheets, often find it vital.

Agree - thats why you can still find Lotus 123 users around. For statistical calculations gnumeric is the superior spreadsheet.

Cons
# High licensing costs Microsoft Office licenses

Sure.

# Advanced features require latest versions

Sure - you need to upgrade your whole office if one person is using that feature.

A bigger issue is that that the XML format that Office uses is proprietary. Opening it in anything other that Office could put you in jail.

I believe I forecasted this a couple of years ago....

A more recent openoffice 2.0 review I wrote

Decline of the American Empire

I spoke with a historian the other night who agreed with my assessment - "All the US needs is 3 more presidents like Bush and we will see a drastic decline in the US Empire".

Bush miss-management has resulted in:
* Breakdown of multilateral organizations (the UN, WTO, kyoto agreement)
* Huge budget deficit increasingly owned by non-Americans (although denominated in USD).
* Increasing distrust and dislike of the US by non-Americans
* Increasing alienation of 1/6th of the world's population - 1.1 Billion Muslims
* Souring trade deficit, and a decline in unaccounted services such as education of non-US university students and tourism due to 9/11 restrictions.

Is military really a US economic competitive advantage?

If you look at the history books, empires that were successes only used military might to open markets (Mongols, UK, Romans, Greeks, etc). Once the market was opened they allowed market forces to take effect. They did not battle with locals to create democracy or extract protection money like the local hood such as the US does in Saudi Arabia.

Domestically Bush hasn’t addressed several important issues:
* Education /re-education of dislocated workers (outsourcing)
* Souring health care costs
* Energy

The most disturbing thing for the US's allies is the muslim issue. Integral to Muslim faith is the feeling that all muslims are brothers. If a brother is in trouble with a non brother it is a muslim's duty to fight for them. Securing every airport is do-able. But securing every subway, train, sporting event, mall, university campus, and public event from terrorism would wreak economic and political havoc.

Productivity gains would fall by the wayside as people would be increasing be focused on personal safety.

Intro to mod_python

Finished an introductory article "The Other LAMP" - on using mod_python. Modpython provides the ability to execute Python code in-process by Apache.

Standards Rant

"Our product utilizes open system standards inherent to Microsoft® products such as COM, OLE, DDE and MTS."

If you see a vendor that says this you should challenge them. Which standards body approved it? IEEE, ISC, ISO?

Here is a list of standards organizations. If it hasn't been approved by one of them then it is not a standard. This is false marketing - I recommend that you remove it or you could wind up in court.

invention != marketplace innovation

Interesting article about invention vs innovation at mitreview.
Samuel F. B. Morse didn’t invent the telegraph. Alexander Graham Bell didn’t invent the telephone. Thomas Edison invented neither the light bulb nor the movie camera...

The simple truth is that the economics of invention are profoundly different from the economics of innovation. Being “first to file” has nothing to do with being first to market. Being first to market has nothing to do with being first to profitability. Being first to profitability—and this is key!—has virtually nothing to do with how quickly, deeply, and ubiquitously an innovation spreads. In other words, there is no meaningful correlation—let alone causality—between a “successful” act of invention and a “successful” marketplace innovation. None.

Why should this surprise us? The marketplace triumphs of “winners” like Edison and Marconi were less a function of inspired inventorship than of ruthless
business practice. .

... Patent productivity is a lousy measure for innovation that matters.

The article made me want to read something by Otto Neugebauer. Unfortunately the local library only turned up a book on the history of astronomy. Amazon wasn't much better...

Cringley good article

Cringley is on his pulpit

"The central point was that paying too much attention to Microsoft simply allows Microsoft to define the game. And when Microsoft gets to define the game, they ALWAYS win."

...intuit...
"Intuit, on the other hand, was suckered into a near-merger with Microsoft that I'm told by some who were there on both sides was only intended to identify and siphon-off the best Intuit technical and marketing people. Through this loss of talent, which was critical, Intuit had no choice but to start competing with its almost-partner."

And some stuff on how Google has commoditized IT servers. During which they have invented a new replicated file system.

Idea!? Take a typical IT function - email, hr, sales force automation - and do what google has done - make it a commodity. I've met with a few companies doing that in retail IT services and currency exchange. Few IT employees, all open source software, and internal IT departments and competitors using propreitary software unable to compete.

Sun Microsystems days are numbered.

Sun has traditionally been a platform (OS + Hardware) for hosting large RDBMS database systems. The database systems that were typically hosted on it were Sybase and Oracle and ERP systems such as SAP. These have been systems that are hard to split up. Typically they scale up rather than out. MainFrame systems for example are typically used as database servers. They called their systems “Open” because at the time the majority of hardware and software platforms (IBM, DEC, etc) were proprietary and closed.

This market is rapidly shifting and Sun - with its high cost structure seems unable to innovate itself out of a market that is being hit with two disruptive technologies, Linux and scalable X86 processors, and one monopolist, Microsoft, who would like to take the RDBMS market. The current three strongest database vendors are IBM, Oracle, and Microsoft and none of them offer any hope to Sun.

Sybase
A while ago Sybase was synonymous with Sun systems in the financial industry. Sybase has been loosing market share ever since around 1994, when Microsoft bought a copy of the source code of Sybase SQL Server and then went on its own way. Sybase has recently shifted its focus from the large server to the mobile market.

Oracle is getting its margins killed by Microsoft and IBM. Its fighting back by using parallel servers which allows you to split up its database onto many commodity systems running Linux. Saves money for the customers and there is more potential margin for Oracle. Oracle has moved its "whole business" over to running on Linux on X86 systems.

SAP is another company weary of Microsoft - they have shifted all their development all over to the Linux platform and are promoting it to their customers. Linux typically runs on Intel or AMD systems.

Microsoft is the fastest growing Relational Database vendor (along with MySQL perhaps), eating market share from everyone and with a huge war chest and aggressive licensing packages to force companies to purchase their database system. Microsoft is essentially giving its software away to large companies at $300-$2000 a pop.

IBM's DB2 typically runs on IBM hardware with AIX, OS400, OS390 or more recently Linux as the operating system. No potential sales for Sun here.

Linux - with the 2.6 kernel Linux now has most of the features needed for large scale relational database hosting. Linux already scaled well to 4 CPUs. Now it with the 2.6 kernel it will scale well in the 16-64 CPU range – Sun’s bread and butter.

Intel – relentlessly seeking higher margins Intel and AMD are scaling up their servers as they gun for the larger profits that the RDBMS hosting business offers.

Herman Kahn 40 years later

I've been reading Herman Kahn's "In the year 2000", a look forward to the year 2000 written in 1965. It is instrumental to see what works in forecasting and what doesn't. Amazingly he got all the major trends right. Forecasting cell phones, satelite tv, the internet, home PCs, dropping birth rates, the rise of Japan, aging society, and the information society etc.. This raises the question, what did he get right and why? What did he get wrong?

I turns out that 80% of the forecasts in computers and communication were accurate. Paul Krugman estimates, however, that only about a third of all the forecasts materialized: “If you go down the list, you will recognize such things as cell phones, the Internet and faxes. But Mr. Kahn's list contains all kinds of things that haven't materialized. Radical new building materials. Undersea cities. Medical cures for cancer and overweight. Only about a third of what he thought were surefire things have come to pass.”

Another way of looking at it is "Where are the flying cars?" or my personal favorite, human hibernation?

This article lists all the forecasts and analyzes the results.

There are three reasons why many of the forecasted technologies do not exist. There are legal reasons, technical reasons and human organizational scale reasons.

One of the main reasons we have no flying cars is legal. As exhibited by 9/11 airplanes are dangerous and thus need to be severely regulated. A stream of flying cars similar to the LA freeway and 20 car pile-ups are legal chaos.

Some things are extremely technically difficult because they contain so many interrelated systems. Biotechnology is a perfect example of this. We have organisms that very similar to humans for testing (rats) and the tools are not particularly expensive. However, it still often takes 10 years of sustained research for a new biotech drug to appear. Our ignorance is best illustrated by Mad Cow disease (BSE) - how can a protein (a prion) cause other proteins to malform? Cross species - through the digestive system!? Medicine receives more research money than any other field but the returns are slow because the systems are so complex.

Finally most innovations that occur can be undertaken by less than 200 people for Logistic reasons. Technologies that require more than that seem only to be created under war effort (Nuclear weapons, Moon Rocket) or totalitarian states (German V1 Rocket, Pyramids). Edison only had 75 people working in his lab to produce light bulbs. Once you exceed the 200 people mark management becomes unwieldy and slows innovation. For example ITER, the international fusion consortium has spent years simply fighting over a location. There are multiple examples of this – the Darpa Great Challenge will likely produce more output than hundreds of millions given to one company. Mapping the human genome was another example of how decentralized innovation is better than centralized.

Examples:

Legal:
69. Individual flying platforms
27. The use of nuclear explosives for excavation and mining, generation of power, creation of high-temperature-pressure environments, or as a source of neutrons or other radiation

Technical – overly complex systems
19. Human hibernation for short periods (hours or days)
35. Human hibernation for relatively extensive periods (months to years)
87. Stimulated, planned, and perhaps programmed dreams
48. Physically nonharmful methods of overindulging

Logistic
94. Inexpensive road-free (and facility-free) transportation
99. Artificial moons and other methods for illuminating large areas at night

Economics books to read

When I studied economics as an undergrad student I always felt that something was something wrong with the fundemental contraints used. I still love reading books about it though. Recently I created an Amazon list of nice economics books I've read recently. Included are books by Friedman , Kay, Krugman, Thurow and O'Rourke. I'd like to read this book on debunking economics

Top exporters

Top exporters from the economist:

The masters of global trade per capita are Netherlands and Belgium, and to a lesser extent Canada, and Taiwan.

Places to visit

I noticed on eve anderson's website she lists places she would like to visit:

"Places I desperately want to visit: India, Santorini, Tierra del Fuego, Antarctica, Machu Picchu, Amsterdam, Japan, Gombe National Park (Tanzania), New Zealand"

Here are several lists of places to visit - and my top places.

BRIC economies
Brazil - Russia - India - China
gs.com

Cheapest 21 places to visit
ASIA
Thailand - Indonesia - Malaysia - Laos - Vietnam - India - Nepal

AFRICA & MIDDLE EAST
Morocco - Egypt - Jordan

EUROPE
Turkey - Bulgaria - Hungary - Czech Republic

THE AMERICAS
Mexico - Guatemala - Honduras - Bolivia - Peru - Ecuador - Venezuela

Most interesting places according to Jim Rogers(jimrogers.com)
Tanzania
Burma
China
Samarkand

My List

I've visited most of the OECD and fastest growing countries. I would like to visit the Middle East, South America, and East Africa.

Middle East
Istanbul
Syria
Jerusalem
Jordan - Petra
Egypt - Pyramids of Giza
Egypt - Sahara

East Africa
African safari
Tanzania - Serengeti
Cape Town

Brazil - Rio de Janeiro - (BRIC economy)
Peru - Amazon
Peru - Machu Picchu

======================

Others:
Antarctica
Ecuador - Galápagos
Papua New Guinea Reefs
Angkor
Taj Mahal
Morocco?
New Zealand?
Pucasawa River - Lake Superior Canada
Jasper National Park - Canada
Sail across the ocean

Interest rate manipulation - does it work well?

Typically governments have two main macro levers for boosting the economy: cutting interest rates and increasing government spending (or increasing government debt like bush has done). Do interest rate levers really work effectively?

They effect certain economic sectors more than others: Auto + Transportation, Construction + Mortgages, and Heavy Industry. What happens when these portions of your ecomomy represent relatively small portions? Multiple sectors fairly uneffected by interest rate cuts include: Heath care, Software, Education, Tourism, Media etc.

It could be argued that by prefering interest rate levers over fiscal stimulus the government is giving preferential treatment to certain sectors of the economy.

MS BS

The dark side of Linux from the globe and mail shows pure FUD from microsoft.

"We have viruses because we are popular(not due to poor coding). Once linux is more popular it will have more viruses".

Most money - $$$ in banks and financial companies is stored in Solaris, AIX or Mainframe systems. Currently these systems are the most popular. Shouldn't they have the most viruses and worms then? No - they were designed securely. Applications such as tripwire and snort are bundled with linux so that it too can have extremely high security. Linux includes a firewall in every server - IP Tables. Open source software is patched more quickly and accurately than MS software.

Most attacks from the internet for windows come through email or downloads (kazaa).

Most linux servers come with all you your software already on the cd. Not many downloads are necessary. Most attacks in the linux world come from buffer overflow attacks. The services on linux that are exposed to the web such as smtp mail, pop3, imap, http (apache) and so on are all very popular so that most of the buffer overflows are already known.

The abundance of scripting and Java vs C++ applications makes linux more secure as well. Scripting languages such as Perl PHP and Python are less open to buffer overflows than c/c++ apps.

A nice way of making your systems more secure is to use a non standard CPU in Linux. Buffer overflows are typically CPU specific, so if you want to be more secure run your linux on a Power CPU from IBM/Apple or a cpu from DEC. Moshe Bar mentioned this technique. Basically most hackers will not have that hardware to develop on and so you should be more free from attack. For example the US military runs linux on Apple boxes.

Myths Of Open Source

Nice article tackling the myths of open source at cio magazine

Here is a list of their major issues with open source:

Issue
1. MONITORING TOOLS

ans: Tivoli anybody? vmstat and sar are fine in almost all cases I've seen.
Add rrdtool and you can create professional reports

2. RELIABILITY

New versions of the Linux kernel have improved the operating system’s reliability, but users say they could always use a more-hardened platform. Disaster-recovery options should be expanded, they say.

Ans: Hot swappable CPU is part of 2.6. Clustering is part of Redhat entreprise 3.
Application level clustering is often more important than hardware.

3. SECURITY

Security continues to be a big issue as backers position Linux as a Unix or Windows alternative in business networks. Efforts to steel the operating system, including the National Security Agency-backed Security-Enhanced Linux (SELinux) project, need to be embraced by vendors.

Agreed - security - ACLs etc need to be mainstream.

4. SIMPLICITY

Linux has the reputation of being overly complicated, and Linux users would like to see that image softened.

Simplicity is often in the eyes of the user. Windows registry I would say is more complex than text config files. Linux / Unix configuration hasn't changed that much in 15 years. I would say lack of change means that it is more simple than windows. In windows the sys admins knowledge is quickly out of date.

5. HARDWARE SUPPORT

Users are asking for better support for things such as third-party drivers, printer management and graphic interfaces.

Ans: That is up to hardware vendors. Much like Windows NT 3.51 consumers of the technology need to be diligent about drivers. The situation is better than Sun, Mac OSX and pretty much any other non windows os though.

6. COHESIVENESS

Users don’t want to see Linux go the way of Unix, where vendors created their own proprietary versions that made it difficult to port applications to one from another. A more-cohesive approach would result in a better operating system, they say.

Ans: Linux Standard Base project addresses this issue.

7. APPLICATIONS

Some users running Linux have found limitations when it comes to software deployment and would like to see a broader range of applications supported, both on the server and desktop.

Ans: This is a chicken and egg issue. However almost all non MS vendors support Linux on the server now. The client needs more time.

8. SKILLED DEVELOPERS

Linux is becoming more widely deployed in corporate data centres, but users say that training for Linux programmers is lagging.

This is pure BS. In my experience the best programmers (even in MS) prefer the command line tools such as grep, make etc that are part of linux. The great books by the late Richard Stevens set the benchmark for excellence in documentation. PHP has become the new programmers web development tool of choice passing asp as Java is the tool of choice for a large number of corporate developers.

My experience with Python has proved most of the smartest and most innovative programmers are at least playing with Linux programming tools at night if they are not allowed to during the day. Python, Ruby, Perl and function languages such as OCAML and lisp have large numbers of brilliant programmers. Those who spend Sunday afternoons hacking instead of at the beach are often the best and the brightest. A year ago I walked by the PHd candidates office at University of Toronto and University of Waterloo. Not one Phd candidate was using windows and a few were using Mac OSX. The majority was linux.
9. GUI

Network execs would like to see an easier-to-use and better-performing GUI.

Ans: Again BS - KDE 3.2 matches everything except Windows XP,2000 and BE OS for responsiveness and difficulty of crashing. Given a year or two it will catch MS. X Windows has a bad reputation but may improve.

10. CONSOLIDATION

Instead of having basic tools spread around the operating system, users would like to see them all in one easy-to-find location.

Ans: Have a look at webmin or alternatively "locate toolname"

IT doesn't matter - much

Retail banking increased its information technology investments in the
late 1990s, when productivity in the broader economy was accelerating, yet its own rates of productivity growth declined. The blame lies with a host of IT investments that failed to deliver as planned.

At many banks, for example, pricey customer-relationship-management
tools haven't boosted the profitability of accounts, nor have more complex,
IT-enabled product bundles increased consumer satisfaction. Unlike
investments that streamline specific processes (for instance,
check-imaging technology), most of the banks' IT spending was aimed at
revenue-enhancement activities that have yet to pay off.

The take-away The experience of retail banks shows that investing
in IT doesn't guarantee improved efficiency when generic IT
solutions are applied to support functions or IT simply represents a "me-too"
investment.

http://www.mckinseyquarterly.com/article_abstract.asp?ar=1154&L2=10&L3=51

Feedback from a friend:
D'accord with my BCG benchmark etc. IT was the largest non-interest expense for banks when I checked it 1999. Another very interesting concept is by Paul
Strass(n)er - The business value of IT. Basically, there is NO direct corelation between IT spending and profits, but here is a very strong case for smart
investing (Not just in IT :-)).

Projects

Cleaning up my email account I found this old but amusing image describing projects. Its interesting that open source software products often nails the user requirements so well - because the developers are generally the same people as the users.

How do you compete with China?

Interesting article from Friedman at the NYT wondering how Mexico can compete with Chinese manufacturers.

I would recommend - don't compete with China or India unless you want to do it on tax relief for companies. Mexico I believe needs to lobby washington about agricultural subsidies. They should focus on food, tourism, fashion that requires quick turn around, the film industry, and attracting lower end military equipment manufacturers (which the US will not outsource to China). A huge potential industry is attracting snow birds (retirees) - cheaper healthcare will be attractive for babby boomers who don't have enough put away for their retirements.

""We are caught between India and China," remarked Jorge Castañeda, the former Mexican foreign minister who just decided to run for president in 2006. "We have lost about 500,000 manufacturing jobs. It is very difficult for us to compete with the Chinese, except with high-value-added industries. Where we should be competing, in the services area, we are hit by the Indians with their back offices and call centers. . . . Not enough people here speak English." And that's not all. While China and India each send tens of thousands of students to be educated abroad every year in science and engineering, particularly in the U.S., Mexico sends just 10,000."

Power Law

Strategy and Business has a nice introduction to power laws in physics and economics. Power law is conceptually similar to the 80-20 rule which you may be familiar with. Buchman has also written several books on the topic.

Weblogs also seem to follow power law.

"Physicists have learned that hidden order of this sort comes with a very simple mathematical signature. A power law is a relationship in which one quantity A is proportional to another B taken to some power n; that is, A~Bn. Trees, clouds, and fractured surfaces all conform to power laws, as do river drainage basins, fluctuations in Internet traffic, the response of the immune system, and a vast range of other natural phenomena. Surprisingly, power laws also arise in the statistics of events that would seem to be utterly random, such as earthquakes and forest fires. For example, the number of earthquakes that release energy E — a measure of their strength — is simply proportional to 1/E2."

This also applies to companies -
"sociologist Robert Axtell of the Brookings Institution, who has studied the empirical distribution of U.S. firms according to their total sales, has found a strikingly simple pattern: The number of firms having total sales S is proportional to 1/S2. In words, firms with sales of $1 million are four times as numerous as those with $2 million, which are four times as numerous as those with $4 million, and so on, right across the board for firms ranging from tiny news agents up to vast multinationals."

mod_python is excellent

I have been working on a site in mod_python with apache for a friend. The tools are fairly low level, but well written. I'll upload a php to mod_python migration document in the next week or so.

Greenspun on itconversations.com

Phil Greenspun, an interesting "open source thinker" has an interview on http://www.itconversations.com/shows/detail94.html.

Some points he makes that I found interesting:

1) Weblogs were surprising because the technology is so trivial
2) In selling open source products you need to be more efficient than
companies internal IT department. As the publisher of an open source product
companies will typically turn to you for updates because they trust you and it
makes it easier if modifications are merged into the main source code.
3) When trying to selling a book that is available online it is important that the
print is attractive and on nice paper. What they can't get from a laser printer.
4) Tools for web development haven't improved that much since the 90's.