Applied Engineering, Not Fiddling Around in the Lab, Delivers
August 12, 2009
Ars Technica has added a dash of whimsy to its article line up. Consider “Weird Science Sees the Eureka Moments Are for the Birds.” A “eureka” moment refers to the story that Archimedes had an idea whilst lounging in his bath. He then ran down the “street”, allegedly naked, repeating “Eureka.” This is the same fellow who would not follow orders and was killed.
Ars Technica’s John Timmer reported a research report about birds solving a problem via tool use:
As the authors describe it, “We presented four captive rooks with a problem analogous to Aesop’s fable: raising the level of water so that a floating worm moved into reach. All four subjects solved the problem with an appreciation of precisely how many stones were needed.”
When I read this, I thought about the Google. The company has an applied engineering side and a provision for some Googlers to fool around in their spare time at work. The notion of one day a week to pursue pursuits may have been crushed by the peer pressure, but that too is a good myth. Microsoft has a significant commitment to R&D as well. I am not sure what the balance between applied and pure research is at either Google or Microsoft.
Based on the reading I have been doing, Google seems to be more like crows. The engineers bang away at a problem until a result yields data. Google then wallows in the data and follows the washing machine instructions on my mother’s Whirlpool. Wash, rinse, repeat.
Google reminds me of these crows. Microsoft? I don’t have an opinion. I will form one when I see the fruits of the R&D applied to Fast ESP and the Yahoo search technology. Both of these are Microsoft acquisitions and both will be applied engineering problems. Another lab experiment which is almost as important at Rupert Murdoch’s “charging for news” experiment.
Stephen Arnold, August 11, 2009
Google Wave: Perturbing the Data Pond
August 11, 2009
I read Anil Dash’s “What Works: The Web Way vs. The Wave Way” and scanned some of the comments about his write up. I agree with most of the information in the write up. The one point I would add is that Wave is not one thing. Wave is a plastic bag made of Google technology that contains quite a number of features and functions. Some of these can operate as a stand alone service; for example, Gmail, the JavaScript beastie of remarkable capability. Other functions need the Google datasphere to work; for example, the time slicing across different Google functions. In my opinion, until the plastic bag metaphor is considered, analyses of Wave will be in terms of software such as SharePoint, which is too limiting for the Googlers’ innovation.
Stephen Arnold, August 11, 2009
Shocking Search News
August 11, 2009
The Washington Post dropped a bomb shell next to the mine run off pond here in Harrod’s Creek a day ago. The Washington Post reprinted a TechCrunch write up that had been timing out for us. The headline? “Which Search Engine Do You Choose In The Blind Test?”
This search tool strips out all the branding, so you’re forced to really think about which results you like better. And early results showed a much more even distribution than Google’s 70% market share would suggest: Google: 44%, Bing: 33%, Yahoo: 23%. The score keeping feature was removed when people found a way to game it, but you can still run the test against yourself and see which search engine you really like the best. Too bad the one I seem to like will shortly be mothballed. The tool was created by Michael Kordahi, a Developer Evangelist at Microsoft.
I am nervous about quoting this much text from a newspaper, but I thought it was important to remind myself that Yahoo will become Bing.com and that the person creating the test works / worked at Microsoft.
I am still surprised by the fact that when Bing and Yahoo are added together, the combined score is bigger than Google’s miserable 44 percent. I had been drinking the Kool-Aid from various consulting firms that Google’s market share was more than 70 percent in Web search. Maybe these are two different scores? Heck. What difference does that make. For this golden moment in research, Bing and Yahoo have more moxie than the Google.
Stephen Arnold, August 11, 2009
Baseline: A Magazine on a Diet
August 11, 2009
I have saved some back issues of Baseline, a publication aimed at enterprise information technology executives. The January 2005 issue contained what I lovingly describe as “crazy numbers”. What this phrase means to me is an analyst / journalist prepares cost estimates for an enterprise software system. The January 2005 article was “Calculating Costs of Installing a Work Flow Management System”. The analysis appeared between pages 44 and 45 as a foldout. The cost for the system was presented as two numbers and the editorial content showed where those numbers originated. Robert McNamara would have been proud of the zero base approach. The assumptions were not as precise as I would have liked, but I did find these estimated costs most useful:
- The start up applications cost was $1,574,720. I don’t know what this phrase means exactly, but, please, keep in mind these are 2005 dollar estimates.
- The operations applications cost was $174,460, which was about 10 percent of the start up applications costs.
- The total applications costs were $1,749,460.
To this total, Baseline’s analysts added:
- Startup infrastructure costs of $247,160
- Operations Infrastructure costs of $39,320, about one-sixth the total infrastructure costs.
I found these numbers useful to explain why a different and more detailed cost analysis was needed. The reason was the various costs that are not included in the Baseline analysis. One example: where is the provision for indirect costs associated with managing the IT crowd when an outage causes extraordinary expenditures. And, where are those provisions for urgent fixes, opportunity costs when the system fails, and the cost of manual workaroiunds as the workflow system undergoes surgery.
Baseline is in rough seas. Image Source: http://www.yourglobaltravelguide.com/wp-content/gallery/namibia/abandoned-boat-in-rough-seas.jpg
I loved this feature in Baseline; it was my stalking horse to show why and ArnoldIT analysis was needed to avoid another information technology cost surprise.
The publication in 2005 had high production value as well. The charts and graphs were painstakingly illustrated. The magazine had 84 pages, some interesting advertisers like Hitachi who paid for a booklet glued into the magazine itself. The editorial was interesting. There was a good mix of features on old shoe subjects like customer service and some briefer items I found interesting like “Jail Break”, which told about a prisoner who compromised secure information.
The Baseline that arrived on Thursday, August 6, 2009, caused me to ask the question, “Why bother?”
The publication was 34 pages and lacked the features that prompted me to save issues for years. The stories were very similar to ones that appeared in earlier issues. The ads, well, were skimpy. The cover showed an illustation about risk but the magazine came in a second cover that said:
Maybe you were out of own. So we are sending you another reminder to renew your subscription to Baseline.
Nah, I was not out of town. I just don’t find the publication compelling any longer. The feature I liked was long gone. If the publication shrinks, what is the justification for paying for paper, ink, printing, distribution, and the entire non editorial work force. Post the information on SquareSpace.com and call it a day.
I did click on the renew button and I was greeted with one of those number online surveys. I know that much of the data collected in this way is bogus. I used to click on the first choice in order to qualify for a free subscription when I used to read trade magazines. Now I look elsewhere. If I want an opinion, I can find a podcast and listen to a pundit.
I think the trade magazine sector may be in for some rough seas. Those numbers were crazy but they helped me sell more sophisticated analyses. I suppose creating those flights of fancy, like the nifty graphcs, were too expensive. Like those professionals, I fear I am a gone goose from the Baseline subscription list.
Stephen Arnold, August 11, 2009
Google Moving Forward in Global Telco Sectors
August 11, 2009
Short honk: Globes reported “Cellcom Launches Israel’s First Googlephone.” Globes reported:
Cellcom CEO Amos Shapiro said, "I am proud to say that Cellcom is the first operator to offer its customers, as of today, the first Android device in Israel. Cellcom is the third operator in the world to launch this device. As always, Cellcom offers its customers the best communications solutions."
The article stated: “Cellcom will charge NIS 105 a month over 36 months for the Samsung i7500 and a three-year subscription to navigation content.”
Beyond Search is not sure about the “first” because India and China may have beaten Israel to the punch. Slow but sure for the Google. A happy quack to the reader who alerted us to this news item.
Stephen Arnold, August 11, 2009
BlackBerry and Quality
August 11, 2009
I bought an 8320 BlackBerry to test the device. It worked until the ball stuck. I got a replacement. I crushed that one and I had to buy a replacement. Now I have a BlackBerry 8900, and it contained information that urged me to navigate to the BlackBerry App Store to buy applications. I have several mobile devices, so the goslings and I went through the normal routine: download the desktop software, update the desktop software, update the firmware on the device, have lunch, take a nap, and then come back to explore the mobile device. I am baffled about [a] the download times and [b] the lengthy firmware update processes. Other aspects of the BlackBerry baffle me as well, but my focus is on the issue of quality control.
The goslings and I picked an application called Platinum Solitaire 2, where some information is available from a third party. The reason was that we figured version two was probably stable and it was expensive for a dorky game like Solitaire. We clicked the download button. We had to hunt around for the icon with a folder and a downward pointing arrow. We clicked the icon. The system promptly froze. We looked at one another. We took out the battery and tried again. Same result. Frozen mobile device.
To reboot this device, one removes the back and removes the battery. Then a long wait as the consumer device figures out that it is what it is supposed to be and that no evil consumer has hacked the phone. I could have a heart attack from the stress induced from this lengthy process. Unacceptable in may opinion.
Now the interesting part of this test was my navigating to the incredibly hard to use BlackBerry customer support Web site, finding the link to point out this problem with the $4.99 application, and wait to see what happened.
Here’s the scenario:
First, I got an automated response the same day telling me that BlackBerry cared. I had to provide details of the problem. I wrote back that I thought the company’s quality control was lousy. I waited.
Second, I got an email from a human who wanted to talk with me. I wrote back that I was working on a blog story about online search for the BlackBerry app store and I did not want to talk to a human. I wanted to use search to solve my problem.
Third, I got an email from another human who provided me with contact information so I could call. If I didn’t want to call the human said, I could expect a call from this BlackBerry person.
Fourth, I got a call on Friday, August 7, 2009, from a really concerned BlackBerry employee. The person wanted me to know that BlackBerry cared about my problem. I explained that I used a new 8900, went through the hoops that BlackBerry puts in front of a customer to get the phone to work, and I went to the BlackBerry App Store. I pointed out that the app called Platinum Solitaire 2 killed my phone.
Google and Real Time Maps
August 11, 2009
A happy quack to the reader who alerted me to GoogleMapMania’s “Real Time Google Maps”. The article contains a number of links to real time Google Maps created by developers. The one that I found most useful was the Chicago Transit Authority map. Google has a burgeoning transportation services business. Those operating bus, rail, and shuttle services may want to take note of this CTA-centric gizmo.
Stephen Arnold, August 11, 2009
Traditional Media Explains Why It May Be Marginalized
August 10, 2009
Short honk: Navigate to the New York Times’s article “Breakfast Can Wait. The Day’s First Stop Is Online”. In my opinion, the article does a good job of explaining why traditional media are being marginalized. Note the pattern. Up. Online. Leave. No time for the newspaper. Not a hint of interest in a magazine. A book. Not a source of fun for anyone except grandmother. How will traditional media respond? Hybrids; that is, a marriage of traditional publishing and new technology. Great idea. How will these solutions be fit into a day that begins with email and ends with surfing YouTube.com?
Stephen Arnold, August 10, 2009
Google Marketing to the Enterprise
August 10, 2009
I usually find Larry Dignan’s view of information technology spot. I was not too surprised with the argument in his “Google’s Campaign for Apps Doesn’t Address the IT Data Elephant in the Room.” The key passage in the article for me was:
In fact, nothing in Google’s marketing toolbox—the viral emails, the YouTube videos and the posters you can plaster near the water cooler—are going to change fact that your corporate data is hosted by Google. If Google really wants to entice the enterprise it should have skipped the YouTube videos and allowed companies to store some of their own data.
I agree that Google has not done a good job of addressing the “Google has your data” argument.
Google has some patent documents that describe clever ways to have some data processed by Google’s systems and other data on a client’s servers with the client retaining control over the data. I am sitting in an airport at 4 50 am Eastern and don’t have access to my Google files. My recollection is that Google has been beavering away with systems and methods to provide different control methods.
The problem with Google is the loose coupling between the engineering and marketing at Google. The push to the enterprise strikes me as a way to capitalize on several market trends for which Google has data. Keep in mind that Google does not take actions in a cavalier way. Data drive most decisions, based on my research. Then group think takes over. In that process, the result is a way to harvest low hanging fruit.
After some time passes, engineering methods follow along that add features, functions, and some robustness. A good example is the Google Search Appliance. In its first version security was lax. The present version provides a number of security features. Microsoft uses the same approach which has caused me to wait until version 3 of a Microsoft product before I jump on board. For Google, the process of change is incremental and much less visible.
My hunch is that once Google’s “go Google” program responds to the pent up demand for more hands on support for the appliance, Apps, and maps—then the Google will add additional features. The timeline may well be measured in years.
If a company wants to use Google technology to reduce costs now and reduce to some degree the hurdles that traditional information technology approaches put in the way of senior management, the “go Google” program will do its job.
Over time, Google will baby step forward. Those looking for traditional approaches to enterprise software will have a field day criticizing Google and its approach. My thought is that Google seems to be moving forward with serious intent.
I think there will be even louder and more aggressive criticism of Google’s new enterprise push. In my opinion, that criticism will not have much of an impact on the Google. The company seems to be making money, growing, and finding traction despite its unorthodox methods.
Will Google “win” in the enterprise sector? I don’t know. I do know that Google is disruptive, and that the effects of the disruption create opportunities. Traditional enterprise software companies may want to look at those opportunities, not argue that the ways of the past are the ways of the future. The future will be different from what most of us have spent years learning to love. Google’s approach is based on the fact that customers * want * Google solutions, particularly applications that require search and access to information. That is not what traditional information technology professional want.
Stephen Arnold, August 10, 2009
SAP and Its Evolving Business Model
August 10, 2009
First, the Tibco rumor and now the SAP on demand software strategy. Managing Automation’s “SAP Unveils an On-Demand Software Strategy for Large Enterprise Customers” surprised me. I had dismissed SAP’s cloud chatter as fog. Not so according to Jeff Moad, a member of the Managing Automation editorial staff. He wrote:
At a recent SaaS conference in Amsterdam, John Wookey, SAP’s executive vice president for large enterprise on-demand, said the company plans to roll out a series of SaaS products for large enterprises that integrate tightly with SAP’s on-premise Business Suite and run on the Java-based on-demand platform that SAP acquired along with Frictionless Commerce in 2006. SAP will concentrate on selling the SaaS offerings to existing users of its business suite rather than new accounts, Wookey said. SAP’s SaaS offerings for large enterprises will include some existing and some new products. Existing products include SAP’s CRM on-demand and e-sourcing services. SAP CRM on-demand will be migrated to the Frictionless on-demand platform…
You can get a consultant’s viewpoint and some verbiage from SAP top brass. For me, the article triggered three thoughts:
- How will SAP make up the shortfall between the revenue from its traditional approach to licensing and deploying its software and the “cloud” model?
- If there is strong uptake for SAP cloud services, from where will the engineers needed to service the clients come? If SAP trims down the functionality, won’t the savvy buyer look for lower cost cloud options or just fire up some coders to create a solution using Google or Microsoft functions?
- What will customers be getting? Will this service be a 2009 version of Microsoft’s early push into cloud computing?
- Whither TREX search?
I don’t have answers, and I didn’t see them in the Managing Automation write up.
Stephen Arnold, August 11, 2009
-
- Subscribe to Beyond Search
Feature archive
News archive
-
