Friday, December 21, 2007
Today I received an email from Spock for someone trying to add the company recruitment email to his "Web of Trust". Fine people harvest email address all the time. What annoyed me is that it required two entries of the email address and clicking on a link in an email.
Not good. Made the job of unsubscribing time consuming. And it wasn't like I was a registered and the email was from someone who had found me but rather spam contacts email. Long story short, do not use Spock. If this what the company puts you through just to unsubscribe from contact spam then I hate to think how annoying the service would be if you are a registered user.
Point to all companies. Unsubsribe must be a single action on the part of the user. Double, triple, quadruple actions are a no-no. And it musn't take 10 days to filter through your system. That is a load bullshit. All it means is that you can't be arsed to fix you email marketing system and you are going to try to spam me as much as possible in that 10 days.
Tags: Spock, Email Marketing,, Email
Monday, December 17, 2007
Several fair comments on from my previous post on using Eclipse/Python so I think I should elaborate on what I see as problematic with Sage. Bearing in mind that I have used Matlab primarily and I am looking for a reasonably priced math program for non-technical users.
Lets first consider what Matlab and its ilk are. They are IDEs for development mathematical-based programs. Most include interactive UI for doing simple calculations. They also provide a large collection of common numerical computations that the user can string together into a larger program for completing analysis of information. So in all there are three elements: an IDE, a math engine and an interactive environment.
Sage is not a very good IDE (actually, I'll re-state that, it doesn't have an IDE). It strikes me as a good program for doing math but it really falls down without an IDE. Given how iuseful IDEs are for writing programs, this is surprising. Particularly when the program is pushed as an open-source replacement for Matlab et al., which do have useful IDEs.
One commentator asked what I found problematic. Well, you know the first one, No IDE. The second was after downloading SAGE, I found I then had to down VMWare, which I could only get after jumping through hoops at VMWare (not particularly happy about that). Third, the notebook interactive UI in the browser is a neat idea but ultimately, it left me underwhelmed.
Now I freely admit part of the problem was that I was expecting an IDE, with an interactive mode and a wonderful math engine. Unfortunately, that was not what I got but that was how it was marketed specifically "a viable open source alternative to Magma, Maple, Mathematica, and MATLAB". It is not an alternative to those programs while it lacks an IDE. The reason people use Matlab is it is easy to use. I really think this is a case of failing to see the forest for the trees. Matlab et al provide more than simply a math engine.
My suggestion to the SAGE team is to ditch the current UI method and use Eclipse. Build a wonderful math engine that uses the Eclipse to provide both a wonderful IDE and an interactive UI. The advantage for the SAGE team is they get to focus on creating an excellent math engine that is open source, while leveraging the work of other teams making Eclipse the best IDE around. Isn't that one of the key advantages of Open Source?
Short term, make it easy or at least write a very good explanation of how to call the SAGE engine from Eclipse.
Granted, I may not understand what SAGE is for, but when presented with marketing "viable open source alternative to Magma, Maple, Mathematica, and MATLAB" I certainly expect something that matches Matlab, if not exceeds it. SAGE does not achieve this.
Tags: Matlab, Mathematica, Eclipse IDE, Open Source, Sage Math, Python
Friday, December 14, 2007
Recently Sage was release to compete against Matlab, Mathematica et al. I had a quick try of it but back out. It quickly became difficult to do much. While I am sure it is very powerful math program, its got a long, long way to go before it offers serious competition to Matlab.
I think a better Matlab-killer is the combination of Eclipse & Python. Eclipse provides the visual IDE of Matlab and Python is an excellent scripting language that can easily replace m.files. All that needs to be done is bundle plotting into Eclipse (something it should have anyway), toss in a large library of numerical and symbolic python functions (complied or C) and you have a program that can successfully compete against Matlab and Mathematica.
Having been essentially using Eclipse and Python this way for the last few days I am convinced it is a very workable solution. Why re-invent the wheel?
Tags: Matlab, Mathematica, Eclipse IDE, Open Source, Sage Math, Python
Web Next is not some quantum leap in reality but rather the culmination of several long term trends. Web 1.0 was the translation of real world services (e.g. Amazon) onto the Internet and Web 2.0 is Darwinian evolution of UI and social media tools and technologies. Web Next is exploiting of information to achieve new products and services with no direct analog in the real world. Some will call this Web 3.0 but I prefer the simpler moniker Web Next.
Web Next is about the creation of value not through the control of information but via the creation of synergies and knowledge through combining information and functionality. There already exists the primitive examples in the Web2 world, those such as the map-based mash-ups. Essentially value is derived via the showing a spatial relationship between data. However, these mash-ups are relatively primitive. They rely on a users existing knowledge of the spatial area in question. I personally have no appreciation of the real layout of New York City having never been there. Consequently, the value I gain from viewing or using a mash-up consisting of crime statistics plotted on a map of NYC is less than someone who has visited which will be less than a resident of NYC.
The synergy of information and functionality is created through Data Ecosystems.
A Data Ecosystem is two or more different data sets that when combined with complementary functionality produce multiplicative effect in usefulness. Or put another way, a Data Ecosystem contains more than one source of data (a data set) (e.g. temperatures and rainfall) that can be combined, analysed and processed with the overall Data Ecosystem being more valuable than data or functionality on its own. Importantly, having more than one data set is not sufficient on its own. Rather you need various tools and functions that allow the user to act on the data. An example will help to clarify.
Take an individual piece of data, say a series of temperatures measurements. On its own you can't do much with those temperature measurements but combined with rainfall measures, annual growth rates and a map, those temperature measurements suddenly have a lot of value. Now a farmer can research and plan when to plant his crops or adjust his crop forecasts based on historical growth rates versus temperature and rainfall. To be able to make the forecast of crop tonnage the farmer needs a series of tools that allow him to find correlation factors and extrapolate the growth trends based on rainfall and temperatures. Without those functions having the data is not particularly useful. There is no use in having gobs and gobs of data if there is poor functionality in the Data Ecosystem.
Data Ecosystems highlight a very interesting point about data. One that I find is continually ignored or not understood by most data companies (including web companies). Data on its own has little intrinsic value. Data only has value with what you can do with it and what you can do with it is determined by what other data you have along with the functionality you can apply to the data.
Like a biological ecosystem, a data ecosystem must mesh together. A Data Ecosystem needs to be internally consistent. If a Data Ecosystem is not consistent then it will not generate value for the user. There is no point in trying to create a Data Ecosystem that has rainfall patterns from Australia and crime statistics in New York City. Designers of Data Ecosystems must not design the systems so they become inconsistent. Consistency is crucial. But given human nature I fully expect consistency will be ignored.
"And there's the sign, Ridcully," said the Dean. "You have read it, I assume. You know? The sign which says 'Do not, under any circumstances, open this door'?"
"Of course I've read it," said Ridcully. "Why d'yer think I want it opened?"
"Er...why?" said the Lecturer in Recent Runes.
"To see why they wanted it shut, of course."
-Terry Pratchett, Hogfather
At this point I expect some readers will be thinking that Data Ecosystems is simply the Semantic Web. Data Ecosystems is not the Semantic Web. Semantic Web technologies will be a part of Data Ecosystems, but Semantic Web is neither a precondition for nor sufficient on its own in order to build Data Ecosystems. Semantic Web helps by automating building the relationships between bits of information. In the temperature example above Semantic Web would have provide information such as the lat/long of the measurements, how it was measured, the accuracy of the measurement, the date and times of the measurement etc. This would then allow a computer to match the data automatically with rainfall data from the same location and time and plot together on a map. Semantic Web makes building Data Ecosystems easier and like objects in programming will allow Data Ecosystem platforms to increase the ease the deployment of Data Ecosystems.
Data Ecosystems can also be built of other Data Ecosystems. The output of several Data Ecosystems can be used as the sources for another Data Ecosystem and so on. Each step creating more value by allowing an individual to achieve more. Data Ecosystems will in effect create an L-space
Why are Data Ecosystems Important?
Data Ecosystems are important for one very, very crucial reason. Data Ecosystems allow people to achieve things effectively. Unlike Web 1.0 which was essentially removing transaction costs from existing real-world processes, Data Ecosystems unlock the potential for new services that are impossible in the real world.
Consider a Data Ecosystem based travel service. Such a service will allow you to research a holiday; book all transport, accommodation and activities; create a comprehensive itinerary of the holiday, send alerts at key points along the trip; calculate how much money you'll spend on the holiday; help you automatically tag video, audio and photos from the holiday and create holiday memorabilia from the items you have uploaded. All through a single Data Ecosystem.
And there are hundreds, thousands, millions of probable Data Ecosystems that have no analog in today's web.
How Things are Already Changing
To close out I want to consider something that has been banging its way around the blog-sphere and offline world: the fate of journalism. Without re-hashing the debate you can read Bill Keller's speech with Jeff Jarvis's responses here and here as background.
If everyone is a citizen journalist, then what is the point of professional journalist? A seemingly valid question but one that has the implicit assumption that both are or will be doing the same process. From the perspective of Data Ecosystems the job of a professional journalist becomes very different from a citizen journalist. The citizen journalist is a source of data. They will most likely only provide a very narrow bit of data on any particular story. Put another way, the citizen journalist becomes a source like the news wires.
The professional journalist moves on from being the source of the story to gathering all the disparate bits of information about a story and then assembling into a consistent and cohesive context around the core story. Professional journalists go form being the gate keepers to information to value builders by creating context to stories. The role of a newspaper/media company is to provide or create a Data Ecosystem within which the professional journalist can assemble, create and publish the context to stories. Within Data Ecosystems, professional journalists, news agencies and citizen journalists will co-exist and combine to produce a more valuable service than either would on their own or exists today.
The media world is already going through pain as it is forced to adjust to the realities of an information economy. Data Ecosystems provide a means to effectively adapt to the information economy. But Data Ecosystems are not limited to media. Data Ecosystems will exist right across the information world. In fact they will reach into material world as L-space is linked into materials at the molecular level. This is the true revolution of Data Ecosystems, they facilitate the merger of L-space and Real Space.
Tags: Web Next, Data Ecosystems, Web 3.0, Web Services, Semantic Web, Web 2.0, L-Space