March 31, 2004

Aberdeen survey of OpenSource dbs

Interesting Aberdeen survey of Mysql, Postgresql and Berkeley DB.
Its available from Sleepy Cat

Points to take home.

1) Most of the users of Postgresql, MySQL, and Berkeley are very happy with their application
2) Most of them do relatively simple queries
3) They didn't mention MySQL MAX

Features missing in open source databases according to Aberdeen:

MySQL does not offer the following key features typically provided in enterprise
databases(according to Aberdeen):
• Incremental and online/parallel backup/restore
• Encryption (security)
• Bit-mapped indexing (for large data warehouses)
• Single GUI administrative interface
• Views
• Stored procedures and triggers
• Object (complex data) table or data-type support
Users interviewed indicate that MySQL is adequately robust in small and mediumsized
applications.

Features desired according to users:
Triggers
Improved Replication


PostgreSQL does not offer the following key features typically provided in enterprise
databases(according to Aberdeen):
• Incremental and parallel backup/restore
• Encryption (security)
• Deadlock detection
• Row-level locking (typically required by large packaged applications) —
although PostgreSQL does offer alternative sub-table locking schemes
• Bit-mapped indexing (for large data warehouses)
• A single GUI administrative interface
• View update/insert/delete

Features desired according to users:
Gui control centre
Improved Replication
Two-phase commit

I think I want to check out the newest version of berkeley after reading this article. I was pretty sure that
NETAPP offers incremental backup of Postgresql - but perhaps I'm wrong.

Posted by Anthony at 05:16 PM | TrackBack

March 30, 2004

impatient for knoppix

Knoppix 3.4 with the 2.6.x kernel and KDE 3.2 should be shipping shortly according to knoppix.net.

Posted by Anthony at 12:42 AM | TrackBack

March 26, 2004

Darpa wrap up

There is a fairly good wrap up of the darpa event over at http://www.xlogs.net/. One thing I didn't know was that the contestants were forced to go at 25 miles an hour. Clearly they would have gone further if they be allowed to go slower.

Most of the robots used some combination of radar, ladar (a laser used to measure distance), stereovision, and the global positioning system (GPS) for guidance. GPS can tell the robot where it is, but not much else. Radar picks out the moving obstacles (other robots, or perhaps real traffic); lasers scan the route ahead, assessing the terrain every few feet. Stereovision on the move is not a simple problem, and sensors with a "steady cam" effect can help when things are bouncing around.

Unfortunately while most of these things work ok at slow speeds they have a tendancy of not working to well at higher speeds.

http://www.plastic.com/article.html;sid=04/03/18/21331292

Posted by Anthony at 01:46 PM | TrackBack

March 19, 2004

Our energy future

Interesting webcast from Richard Smalley(Nobel laureate) about the future of energy. Also given to another group in Realplayer.

Points:
* Over 80% of all energy currently comes from Oil, Coal and Gas
* Billions of people in the world need more energy (China, India etc)
* Fossil fuel supply (except coal) will significantly decline in the next 50 years
* Any solution that doesn't have the potential of 1 Terawatt isn't a solution
* Nanotech offers some solutions (transmission on the grid, batteries, solar efficiency).
* It takes 20 years to make a top physicist/engineer. The US isn't doing it - Asia is.
* Believes in a distributed grid solution (like Lovins)
* Batteries need to improve by x10-x100 times
* Solar cells need to drop in price by 100
* Use the moon as a giant solar collector?
* Is there one solution? We don't know so we need to try many things
Fusion, Solar, Geothermal, 2nd Gen Nuclear

oil1.png
oil2.png
oil3.png

Posted by Anthony at 04:42 PM | TrackBack

March 17, 2004

Hybrid cars taking off


Nice article from technology review regarding the future of hybrid cars and manufacturers current plans. Hybrids can save $5000 in GAS over the life of the car and could reduce US foreign oil dependancies.

Toyota seems to be the company to beat. The most complex part of the technology is the batteries - life maintenance etc. Capturing the lost energy from breaking is more trivial, the electric motor switches roles and serves as a generator, using the car’s kinetic energy to recharge the batteries. Still, "When it comes to engineering the system as a whole ...Toyota has three, four years’ advantage over the others". I believe US firms will catch up if they want to. European firms (VW) seem to be focused more on Diesel and biodiesel.

DaimlerChrysler
Began fleet sales of mild-hybrid diesel-electric Ram pickup in 2004; experimenting with full-hybrid diesel-electrics

Ford Motor
Canceled Explorer SUV hybrid introduction in 2001; full-hybrid Escape SUV now due out this summer; full-hybrid option promised for Futura sedan, which debuts in 2005

General Motors
Canceled full-hybrid VUE SUV promised for 2005; mild-hybrid Silverado and Sierra pickups to debut this year, the VUE in 2006; promising full hybrids, including a Malibu sedan and Tahoe and Yukon SUVs in 2007, and Silverado and Sierra pickups in 2008

Honda Motor
Insight two-seater, a mild hybrid, was first hybrid to reach the U.S. market in 1999; now sells Insight and mild-hybrid Civic worldwide; adding mild-hybrid Accord sedan this fall

Nissan Motor
Test-marketed hybrid Tino compact car with proprietary lithium battery in Japan in 2000; licensing Toyota technology for a possible 2006 full-hybrid Altima mid-size sedan

Toyota Motor
Launched first hybrid car in 1997; sells full-hybrid Prius worldwide and six other hybrids in Japan; full-hybrid Lexus and Highlander SUVs promised for export by fall 2004 and early 2005; considering full-size hybrid pickup trucks and Camry

Posted by Anthony at 09:09 PM | TrackBack

March 16, 2004

Thinking about energy

Interesting webcast about energy from quirks and quarks. The better interview was with Vijay Vaitheeswaran, author of Power to the People. Here is another more in depth interview with the Author. And kicking the Oil habit on MITs website.

The major points that he makes
* Oil was forecasted to be $100 a barrel by 2002 by the US government and Exxon. What happened? Technology improved.
* 25% of world reserves are in Saudi Arabia. Iran, Iraq, Kuwait, and the United Arab Emirates-each have about a tenth of the world's reserves. Taken together, then, this volatile corner of the world sits atop two-thirds of the world's oil.
* The government needs to cost in externalities into Oil prices
* Governments should stop subsidizing oil prices (Iraq)
* Number one cause of environmental damage is hydrocarbon usage
* Many of the other environmental problems we face (water etc) all are solvable with adequate energy
* The developing world all would like use more energy (transport, cooling, lighting etc).
* He is very pro Hydrogen
* Small is beautiful (Amory Lovins). Smaller distributed groups will drive the solution. Not government or large companies.

Posted by Anthony at 01:12 PM | TrackBack

March 12, 2004

American Job Growth Jitters

Why are American's so nervous about their limited job growth? Their unemployment rate is still far below the long term unemployment rate in Canada , the EU and indeed most of the world.

They press seems determined to blame offshoring [Tom Peters] when the real cause of the jitters are high population growth (driven by the million immigrants a year) and the lack of a decent safety net for those without work. In particular healthcare and job retraining programs are areas that need work and I don't think the Bush Administration is willing to look at.

Posted by Anthony at 09:59 AM | TrackBack

March 11, 2004

How many levels of management in your MNC?

I watched an interesting webcast of Philip Condit, CEO of Boeing over at mitworld. It was filmed in 2003, but seemed very 2001 - ie very pro Information Technology.

Conduit argues that we are only at the begining of the information technology revolution, which much like the industrial revolution, is extremely disruptive, forcing people to change their lives or perish. About half the US gross domestic product is defined by transformational activities, such as making airplanes or hamburgers. The other half of the economy is defined by transactional activities—information passed back and forth among people. Technology is creating a vastly more efficient transactional economy and we are just at the beginning. Condit cites studies showing that faster computing increases productivity, while drastically decreasing the workforce. Condit’s conclusion: We’ll need to rethink the organization of our institutions to ensure that the current economic revolution poses an opportunity as well as a challenge.

Some interesting ideas:
Large companies will be able to reduce the numbers of layers of management due to communication improvements. Boeing has 8 layers, but he sees the future as being more like software companies where corporate heirarchy is extremely flat.

I would like to see the numbers of layers of management tracked in annual reports. I believe it is a better indicator of productivity of a firm than the number of FTE.

Employees and customers will increasing be located anywhere in the world. But the need for travel will remain.

Posted by Anthony at 01:40 AM | TrackBack

Productivity gains driven by IT

MIT Technology Review has an article analyzing recent productivity gains in the US economy. It argues the gains (8%) are primarily due to the infrastructure put in place over the past 5 years (supply chain software, ERP etc). Brynjolfsson is over simplifying things. Thurow is more accurate with his three pronged cause - Information Technology, Market Pressure (shareholders), and outsourcing (lower productivity jobs moving out of the US).


From the 1970s into the 1990s, U.S. labor productivity grew by barely 1.4 percent a year. Many economists thought it would be stuck at that level forever. But the growth rate jumped to more than 2.5 percent in 1995 and has averaged more than 4 percent since 2001.The difference is dramatic. It takes 50 years for living standards to double if productivity grows at 1.4 percent per year, but only 18 years to double at 4 percent growth.

The productivity boom is rooted in a revolution in the way American companies apply information technology. Technology-driven innovation is reshaping the economy, but managers who sit back and wait—assuming that technology alone will quickly or automatically introduce gains—are setting themselves up for failure.

....
The unsung heroes of the IT revolution have not been the microchip and the Web browser, but rather the creative, diligent, and painstaking work done by those who have been rethinking supply chains, customer service, incentive systems, product lines, and 1,001 other processes and practices affected by computers.

Posted by Anthony at 01:07 AM | TrackBack

March 09, 2004

Countries visited

Cool website which allows you to chart which countries you've been to.


If you visited a country before they broke up (eg Russia) do you include the new countries? I didn't.

Posted by Anthony at 01:46 PM | TrackBack

March 08, 2004

Wireless VOIP Pager Phones

Part pager, part IP Telephone these little devices from Vocera Communications are helping nurses get their jobs done.

It is windows only, and uses 802.11b. They are small enough to pin on your uniform.

Posted by Anthony at 02:33 PM | TrackBack

March 05, 2004

Microsoft Dumping?

Strikes me that Microsoft is dumping some of their technology below cost.
In particular the IIS, .NET server and SQL Server.

Corporate price - SQL Server for $300. Windows Server? $200. Makes it hard to sell Linux and/or J2EE. Its cheaper than MySQL corporate edition. Prices for SQL server are a bit more if you want to put it on the open internet - but not much.

Posted by Anthony at 10:45 PM | TrackBack

More on zope & plone

Plone 2 is ready to be released any day now as is Zope 3. The programming model of Zope 3 is supposed to be far simpler. Spoke with a developer last week about zope. He said they have ripped out parts of zope - sessions, zodb and a few other parts but are generally not using it. ACS is the other open source CMS contender - but it is really more of a traditional web database system.

Posted by Anthony at 10:40 PM | TrackBack

March 03, 2004

Looking at Plone

I have been looking at Plone/Zope as a possible replacement for Lotus Domino.

It reminds me of what might happen if Lotus open sourced Domino. Tons of nice templates and apps - but lacking documentation.

For a Domino person the Object Oriented stuff is ok to wrap your head around. All the templating languages are a bit confusing though.

1) I installed the plone + zope download

http://plone.org/downloads

Easy - but what about plone 2? They recommend it for new dev work but don't include it packaged on the downloads page.

Created the admin account
Created 2 user accounts

2) Logged into the manage view
http://127.0.0.1:8080/manage
Seems as if you don't have the 8080 you don't get the control panel
Created a couple of foulders and examined some of the options. Everything is an object - like in Notes everything is a Note.

3) Followed the Flash Video Demos here
http://movies.bluedynamics.org/plone_videos/
Logged in as a user
Created document
Created News Item
- set date
- published it
Created Calendar Entry
- there doesn't seem to be mail integration with this....
- I would have thought it should mail an ical attachment

The structured text is very cool. I would like to get the editor working like in the video.


4) The switch views - to contents view
Played around with the basic state tab

The Members view should simply list members if there are less than 50 of them. Searching by name is a headache.
It should be do-able as the demo site does it this way.

5) Worked through the ZopeZoo Example
http://zope.org/Members/jwhitener/zopeZoo_2_6

6) Changed the Portal Icon, Description etc
http://plone.org/documentation/book

7) Changed the tabs on the opening page
http://plone.org/documentation/book/b

8) Tested the issue collector
Looks ok - a bit rough around the edges

9) Checked out http://demo.plone.org
They have some cool addins - the wiki, discussion board, mpoll(polls), HTML Link
Tested copying and pasting between folders


10) Tested FTP
Logged into the local machine via FTP using the admin name and password
ftp://admin@127.0.0.1/

Fairly cool - you can access your documents quickly this way. What happened to the file extensions though? (no .ppt)

11) Backup and Restore
All the data is stored in Data.fs which is located in var directory with your log files
Stop zope
zip Data.fs
restart
backup

Also: A nice explanation

12) Created Formulator Form
Very easy Formulator from the management interface. Haven't figured out how to include in on the site though.

Read article on creating forms with Zope

Another hint

Still need to test
Create a python script and have it called when you click on an Icon
LDAP Authentication against Novell LDAP or Active Directory
Exporting/Importing data via CSV
Workflow routing
User/ACL Management - GRUF
Storing Data in a relational db

Posted by Anthony at 01:07 PM | TrackBack

March 01, 2004

Will & Vision: How Latecomers Grow to Dominate Markets

Finished reading "Will & Vision: How Latecomers Grow to Dominate Markets" and was a little depressed. Statistically companies that create breaktrough products don't do well.

There is no "first mover's advantage". Companies that rely on being a first mover tend to be slaughtered by companies that have business process innovation and are good at copying others ideas (walmart, microsoft).

Posted by Anthony at 03:54 PM | TrackBack

Coming soon from Bill Gates

Its interesting that Gates is going around the US trying trying to get students pumped up on Computer Science. When asked if he would study CS Gates answered "computational biology" or "artificial intelligence". When asked if there could be another Microsoft he said - a breakthrough in AI would be 10x as big as MSFT.

By saying that Gates has basically admitted that the opportunites for big advances in CS for the next 15 years are limited.

To be really good at computational biology (aka Bioinformatics) you are better off with a MS degree in Biochemistry and learning python or perl for hacking.

With AI there hasn't been any breakthoughs in 20+ years when the hot stocks were all AI companies. Progress is slow. Handwriting recognition works ok. Voice/Vision recognition is limited. We are at the cockroach level and need to get to the human level. There are less actual jobs than in the 80s and the programming tools are more primative. (Java + C# vs LISP).

Marvin Minsky the grandfather of AI called AI Brain Dead and that there hasn't been much progress since the mid 1970s. According to him Clippy and OCR don't cut it.

Gates uses progress in hardware as a proxy for progress in computer programming.

* Harddrives with Terabytes of data
* LCD screens are getting cleaner and may be used as ebooks
* Wi-Max (802.16) extends connectivity from 200 feet to 30 miles.
* SPOT watch which receives data via FM

Web services standards are supposed to be the next big thing in software according to Gates. His example was scheduling. When will Microsoft open Outlook/Exchange to webservices? I see proprietary companies loath to open their proprietary APIs.

Personally I think that real opportunities are in "disruptive technologies". Open Source and specifically applying OpenSource to Financial Services. Or perhaps students should just follow the opportunist Larry Elison's recommendation and get into Biotech.

Larry: "This is just the beginning of the biotech revolution. It's going to change our world even more than computers did."

Posted by Anthony at 02:21 PM | TrackBack

86% profit margins

Software for large companies has one of the highest profit margins anywhere - typically 86% for companies like IBM and Microsoft according to a recent Bloomberg article profiling the growing battle between ibm and msft over linux and open source. IBMs services average between 23-28% while Infosys is in the 40% range.

When you look at how much large companies such as banks pay for software such as databases, email and office suites you realize that if purchasing were pooled between companies and the 86% were put into open source the final product would be as good at a fraction of the price.

For example the City of Toronto spent $86,000,000 on Oracle licenses a few years ago. If $73.6 Million were put into Postgresql development (and attracting the developers to live in Toronto) the final product would be something that is 80-90% as good as Oracle for and other cities around the country/world would benefit.

$73.6 Million buys a heck of a lot of Open Source software. $12.4 would buy more than enough development.

Posted by Anthony at 12:16 PM | TrackBack