Beyond Search

Kindle Craziness

2009-12-07T11:21:13Z

For some interesting MBA logic, read “Kindle Fantasies Are Running Wild — But, For Now, Amazon Is Losing Its Shirt.” I am not too keen on Amazon or the Kindle. Amazon’s financial reports are not too informative, and the data about its cloud services are expressed in mythical Amazon units. Spare me. The Kindle is annoying in its various incarnations. Among the reasons is the dorky snap on cover that can break the device and the when-pigs-fly search system that often launches itself when packing the gizmo away when the pilot tells me to turn off electronic devices. I have learned to live with the gray on dark gray screen and the clumsy keyboard, the experimental browser, and odd buttons. Kindle 1 had buttons that were too easy to push. Kindle 2 have buttons that are tough to use when reading on the elliptical exercise machine. The big Kindle is an invitation to breakage. PDFs and graphics generally are miserable experiences.

The article cited above sidesteps my points and goes into a more murky area; namely, the finances of Amazon’s selling eBooks. The key passage in the write up in my opinion was:

Publishers should be able to sell e-books to distributors like Amazon at $5 and still maintain the profit margins they enjoyed on print book sales. In turn, distributors like Amazon should be able to sell e-books at the current $9-$10 price and still enjoy a healthy profit. The bad news for authors is that their royalties will decrease since they are based off of retail sales price.

I think that books of any type are going to be a financial challenge whether in print, online, or on a device like the Kindle. In case you have not noticed, books are not the way folks under 20 get their info. I don’t see the demographic bulges that are bursting with video, Web, and tweet consumers suddenly jumping on the book bandwagon. There are 320 million or so people in the US and most book readers are concentrated above 42nd Street, in university burgs like Charlottesville, Virginia, Washington, DC, and a half dozen other major cities.

The numbers presented in the write up appear rational, but the thinking that gives rise to this type of analysis is oddly disconnected from the reality of the publishing business. Can traditional publishing survive with a shrinking pool of book readers? Check out the argument in Carnival Culture: The Trashing of Taste in America. That 1992 future is now here. When a blockbuster fails to materialize for book publishers, the decline will be sure and steady. In fact, that’s what is happening right now in publishing. When Google provides good enough options to get nuggets of information, the revenue bleeding will be slow and the open wound will lead to financial Streptococcus pneumoniae. Prognosis: no marathons for the victim. Publishing outfits may just wheeze along.

Don’t expect gizmos and price wars to change the descent. Just take a look at your kids. Reading a book? Not likely for the majority of families I assert.

Stephen Arnold, December 8, 2009

Oyez, oyez, I want to disclose to the Department of the Army that books are not what they once were. If you know someone in the Army responsible for recruit training, run your book argument by that non commissioned officer and let me know what you hear. By the way, this is a free blog. You will have to pay the author of the cited article to get their reasoning for financial bonanzas.

Microsoft Academic Research

2009-12-06T15:54:24Z

Microsoft’s answer to Google Scholar is publicly available at http://academic.research.microsoft.com/. I saw the service in 2006, but then I lost track of it. I saw a post in Research Shelf in late October 2009. I fiddled with the system today (December 5, 2009) and found it useful. There are more than three million research papers accessible via the system. For a full run down of the service’s features, navigate to “Here they Come Again? Microsoft Research Launches Academic Search Database (Beta)”. I am not a scholar and some of the functions won’t be of much use to me. My test queries for various Google wizards and technologies were in line with what my team has gathered over the years. Here’s what a query for Ramanathan Guha generated:

The tabs across the top of the page slice the data by authors, conferences, and journals. I looked at the results gathered under each tab, but for my work, these were not particularly useful. The related authors listings were interesting. My particular approach is to read the abstract and look at the list of authors, paying particular attention to the order in which each is listed in the source document.

The system–object-level vertical search research–was responsive but there was none of the value added hyper linking that is now being used in Google’s legal collection.

When I ran a query for “Jeffrey Dean”, I got some false drops and hits that pointed to authors named “Dean” working in fields unrelated to computer science.

My view of the service is that it is useful. I would use the service, but it would not be my starting point. My own collection of Google information struck me as more conclusive. When researching academic content, I need to be able to jump to patent documents by a particular author and access Web log posts by the individual whom I am researching.

The content seemed okay for the technical fields in which i have an interest, but I was not able to determine how frequently the index is being refreshed. I have some newer Google technical papers I downloaded directly from the Google.com repository. Microsoft may want to make some changes to its crawler. I think Google does a better job of indexing Microsoft than Microsoft does of indexing Google.

Stephen Arnold, December 7, 2009

Listen up. I was not paid to write this blog post. I wish to report its status as a freebie to Lead Hazard Control (Housing and Urban Development Department).

Government 2.5: Traditional Information Technology Evolves

2009-12-06T18:56:54Z

I have just returned from my endnote at the International Online Conference in London. On December 14, 2009, I will be taking one of the 10 trends for 2010 from my London UK talk and expanding on the idea of dataspaces, not databases. Most governmental entities are anchored in traditional database technology. Although state of the art in the 1970s, the RDBMS framework is ill suited for the rigors of Government 2.5 information.

I will be attending the CoolBlue Government 2.5 conference in Washington, DC, on December 14 and 15, 2009. You can get full details about the conference from the program’s Web site.

You can get a glimpse of what’s in my talk. Just search this Web log for the term “dataspace”, and you will get some background information. The dataspace technology is one of Google’s crown jewels, and it a core capability little known outside of a small circle of wizards. You can see a tiny fragment of the dataspace technology in action if you navigate to the Google Wave information page and do some exploration.

My remarks created quite a stir in London on Thursday, December 3, 2009, and I anticipate a similar reaction in Washington on December 14, 2009. Googlers are largely unaware of the dataspace technology, how it embraces the Google programmable search engine, and the company’s push to become the Semantic Web.

I will be linking these technologies to likely government use cases. If you want to talk after the event, just write me at seaky2000 at yahoo dot com. I will make time to visit with Government 2.5 attendees.

Stephen Arnold, December 7, 2009

Oyez, oyez, I want to alert the mayor of Washington, DC, that I was not paid to write this blatant self promotion or mention the CoolBlue conference. I think the conference’s PR manager will buy me a Diet Pepsi. I have my Web feet crossed.

Maggwire: A Remarkable Reinvention of a 1980s Database

2009-12-06T15:53:49Z

I saw a reference to the “iTunes of magazines.” I took a few minutes to track down this reference. I located what may be the ur-reference, an interview with Ryan Klenovich and Steve DeWald, two former investment bankers with Deutsche Bank. You can find the full text of the interview in which this phrase appears in Mr. Magazine’s Web log article “Magazine Innovation in Practice: Maggwire, the iTunes for magazines?” Several points in the interview made it into my paper notebook via a goose quill pen. For more insights about this “iTunes for magazines”, read the article and watch the video.

“We’re trying to become an online destination for reading magazines, similar to YouTube where users associate YouTube as the online destination for viewing videos.”
“Initially, the excitement is going to come from the added traffic.”
“You can call us the Google of magazines, where we crawl the web for these great magazine articles, the users go on there to read and rate the articles, and the idea is that the best ones will come to the top.”
“The ideality is to create this marketplace using one ID.”
“Like I said, we’re not going to replace print, what we’re trying to do is compliment the print magazine.”
“Another thing that we have thought about is, after we do reach a certain critical mass, we believe people will be willing to pay a nominal subscription for our service especially as the publishers continue to expand their offerings and partner up with us and our site will just get better and better and better.”
“I think now we’re really in an interesting phase where you have these e-ink devices and iTablets coming out so the hardware will be there, so you need some kind of central destination to back that up.”

Interesting premise.

Stephen Arnold, December 7, 2009

Oyez, oyez. A freebie. I wonder to whom I report this fact. Perhaps to the iTunes of the US government, the National Archives, maybe?

AOL Finds Role Model: Apple

2009-12-06T20:16:02Z

Short honk: This International Business Times’s article “AOL CEO Looks to Apple for Turnaround” is not about search. It struck me as somewhat unusual. I have been out of the country for a week and I am trying to readapt to the American business scene. Apple is a hardware company that has used software to lock in customers. AOL is an online company which seems to have been the inspiration for Yahoo. The phrase that caught my attention was:

Armstrong, a former Google Inc executive, has used Apple’s experience to create his own plan for a turnaround which he sums up in the report as: “New Products and services that people find necessary.” AOL Inc., which dominated the online experience for many computer users in the mid to late 1990s, suffered as customers migrated from its dial-up online service to Internet service providers.

Yikes! Will AOL try to get into the hardware business like Barnes & Noble and TechCrunch? Must have information services are those that make money for greed crazed MBAs on Wall Street, doctors trying to save a child’s life, and lawyers who want to win the big one so a larger yacht can be purchased. AOL is a precursor of Yahoo.

The Associated Press reported in “AOL Ends Ties with Time Warner” that AOL is an “independent company.” Sink or swim time.

I think the former Googler may want to hunt for another role model. A good place to start is with the search experience for AOL users. Then go from there to monetization. I am not sure Love.com will do it for me. Just my opinion.

Stephen Arnold, December 7, 2009

Oyez, oyez, Securities & Exchange Commission. I know you are ever vigilant, tireless in your scrutiny of publicly traded companies, and working on nights and weekends to get the economy righted. Please, take time to accept my admission that I wrote this news item without compensation, inducements, kick backs, promises for future considerations, free travel in corporate aircraft, paid vacations, meals at the Willard Hotel, or any other type of compensatory action. Just like you guys because you are my role models.

The Google Gong Rings for ProQuest and Dissertation Content

2009-12-07T11:20:33Z

A MOVIE CAMERA BUSINESS TRIES TO ADAPT

In June 1986, I was sold along with the electronic publishing assets of the Courier Journal & Louisville Times Co. to Bell+Howell. B+H owned a new media company, which in the late 1980s did business as University Microfilms with the acronym UMI. At that time, the company’s product line up spanned a number of media. At one end of the spectrum was the original business based on creating microfilm replicas of documents. These microfilms were sold to libraries. Generations of students used technology perfected during World War II for their access to information not in a library’s collection. At the other end were the electronic products from the Courier Journal: ABI/INFORM, Pharmaceutical News Index, and Business Dateline (the first full text local business news database with corrections made when the source updated the story’s facts).

Now this is an efficient research tool for today’s student. Source: http://www.archivalsuppliers.com/images/Picture%20284.jpg

When I was a student, I did research with microfilm. It was okay but it took a long time to get the reels, get them set up, and reviewed. Getting a hard copy of a document was a hassle. Some of the prints turned black and became unreadable quickly. I once dropped a reel and watched in horror as it unspooled, picked up dirt, and was unusable. I had to pay the library for a replacement. I think in the 1960s, a single reel cost me about $45 which was more than I made in my part time job. I loathed the stuff.

At the recent Online Information 2009 event in London, my colleague Ulla de Stricker was the keynoter for the “Publishers Delivering Value” track on December 3., 2009. In her talk – which she mentioned the Google move into dissertations. Her ference inspired me to write this opinion piece. You can get iinformation about her at DeStricker.com. One of her exampls was the fact that Stanford University students may now submit their dissertations to Google while it is optional to submit them to ProQuest.

So I wandered over to the exhibit hall to visit with ProQuest, all the while reminiscing about my past experience with that company – known as UMI at the time.

MICROFILM: HARD TO USE, EASY TO DAMAGE AND MISFILE

When I was working on my PhD, I remember my fellow students talking about the costs of getting their dissertations “published” and then included in the Dissertation Abstracts index. I never had this problem because I took a job with the nuclear unit of Halliburton, never bothering to submit my dissertation once I got a real job.

A microfilm readers. Source: http://www.ucar.edu/library/collections/archive/media/photographs/481_1976_microfilm_lg.jpg

The whole system was a money making machine. When a library burned down, backfiles could be acquired when physical copies were not available. When a university got a grant for a new field of study, a collection of journals could be purchased from UMI on microfilm. Bang. Instant academic reference material. I don’t recall how much content the “old” UMI moved to microfilm. My recollection is that there were books, journals, newspaper, and, of course, dissertations. With all this film, I understood why B+H had paid tens of millions for the Courier Journal’s electronic publishing expertise. Buying expertise and using it are two different things, in my opinion.

MECHANICAL PRODUCTION WRONG FOR DIGITAL PRODUCTS

The production process for creating a microfilm was quite complicated and involved specialized cameras, film, and chemicals. The image I have of the UMI facility in Ann Arbor, Michigan, the first time I visited was a modern day castle surrounded by a moat. The castle was a large, one-story building surrounded by a settling pond. The chemicals from the film processing were pumped into the moat in order to separate certain high value residue from other chemicals. UMI processed so much film that the residue silver from the photographic process warranted this recycling effort.

Dinosaurs struggle with the concept of an apocalypse. Adapt or get grilled I suppose.

UMI had a silver mine in its monopoly on certain types of content. My recollection of UMI was that its core product was getting universities to require or maybe strongly recommend that doctoral dissertations had to be “published” by UMI. The microfilm copies of the dissertations were sold back to the doctoral students and to libraries interested in having a compact, relatively easy way to store volumes on a mind boggling range of topics. I did a project that required me to use a microfilm copy of something called the Elisaeis by a wild and crazy religious poet named William Alabaster, and several dissertations written about that nearly forgotten literary work. I also did a project for the Vatican and worked through microfilms of sermons from the middle ages in Latin. Now that was fun! Pretty sporty content to. Nothing like a hot omelie.

I never paid much attention to the microfilm business processes, but I did quite a bit of thinking about craziness of making film copies of documents that could be converted or processed in electronic form.

After a couple of years laboring in the parched fields of Bell+Howell, I jumped to Ziff Communications and focused my attention on electronic products. I recall giving a couple of talks about the differences between information companies founded on mechanical and physical processes and companies focused on more digital processes. The example that I used repeatedly was the UMI thought process. Handling paper to make film affected everything the company did during my brief tenure. The collision of that type of rear view mirror information product and the forward looking digital products made me a casualty of the collision between the cultures of mechanical and digital work methods.

I know that the products with which I had been associated during my brief tenure at UMI have not changed very much. In fact, when I was in London last week, I stopped by the UMI booth—the company now calls itself “ProQuest”, which seemed more appropriate for a golf club company or a vacation planning agency for extreme adventures—for a briefing. I have to admit that I was disappointed. Not only did the knowledge of the sales person disappoint, the core features of the digital products on which I worked had not changed in a material way since 1986. A quarter century of innovation and a logo changed. Interesting.

The reason I am thinking about this subject is my reading in the Stanford Copyright & Fair Use Web log “Stanford Dissertations Moving from ProQuest to Google.” The story is simple: “Stanford is partnering with Google to make student dissertations available worldwide.” The blog post provides some detail about what’s going on. The key point for me was this exchange between Stanford professionals:

Minow: I understand that this move away from ProQuest means that Stanford student work will no longer be included in Dissertation Abstracts unless the student makes an affirmative effort to submit to ProQuest. What are the implications for the broader research world of such a step?

Calter: It is a concern, but our sense is that the wide availability and visibility of the dissertations through the Stanford catalog and Google will more than compensate for the lack of a listing in Dissertation Abstracts.

VALUE OF DISSERTATION CONTENT

Why is this single announcement from Stanford important to me? Certainly not for the reasons spelled out in SFGate’s “Google Publishers Stanford Dissertations Online”.

First, this is the beginning of the end for the “old” UMI monopoly on dissertations. Google can suck in this content, process it in the way it is parsing and tagging its legal content, and make it available under its subsidizing business model. The entire mechanism is simply more in line with what PhDs know how to do with information. I am surprised it took Google 11 years to move into this content space.

Second, the commercial database companies like ProQuest, Reed Elsevier, and Chemical Abstracts, among others, have a clear view of their future. I don’t think the Mayan calendar’s hint of an apocalypse was intended for traditional database publishers, but that end of history date a few years out may apply to the traditional electronic product business sector. Google has the scale and the demographic momentum. A single database company will find it tough to deal with Google’s combination of scale and its subsidizing business models.

Third, I think dissertations are useful in certain types of recruitment and competitive intelligence work. Dissertations are written by individuals with the help of some advisors. Useful details in my experience. Quite useful.

Fourth, the online world has begun to shift rapidly from proprietary, for fee services to the broader “search box” model available from a computing device equipped with a browser. I don’t think dissertations are viewed today as a particularly useful source of information by most users. If Google sucks in more dissertations and creates the type of value added tagging evident in its legal collection, dissertations will become a much more used source of information. I know the value of dissertations. These documents contain useful bibliographies. Some include tabular data not available anywhere outside of the dissertation document itself. Often a dissertation will contain a detailed description of a process that delivered a particular result. Process information or how to content is quite difficult to find in other types of source material. Finally, I have found dissertations a useful resource when researching prior art.

I hope the ProQuest team can adapt to this Google challenge. If it does not, ProQuest may become a case example of a company whose revenue stream was captured by an online service that combined better technology with a business model more in tune with users and the economy.

SOME COUNTER ARGUMENTS TO MY OPINION

I am offering my opinion, gentle reader. I can envision a range of arguments that traditional database vendors will throw back at me directly and indirectly. For example, look at these in the table below:

Counter Argument	My Opinion
Our company does a more professional job of indexing content than an automated system like Google’s.	No. Goggle’s tagging and classification system is either comparable or better. Also, Google’s method is cheaper and faster.
Our databases exist as a family so researchers can find high quality information quickly and without the noise in a Google result list.	Maybe. Google’s content base is large and its system learns. The tagging and linking begin at “good enough” and then move toward “better”.
Google cannot handle graphics like chemical structures.	True. Google has patent documents that indicate that rich media are in the company’s bag of tricks, just not yet put in the public facing products.
Universities trust our indexes and knowledge products.	That’s changing and fast.

I won’t drag into this discussion the technical differences between UMI / ProQuest and Google. These are simply too different to warrant comment.

BEGINNING OF THE END FOR SOME INFORMATION COMPANIES

I think Google has a number of options at its disposal. I think Google will take baby steps toward commercial online content. Who knows? Maybe the Google will stumble and the world will go back to the good old days of microfilm and Dialog Information Services?

I think that after 11 years of viewing Google as a Web search company that sells ads, it might be tough for the commercial database companies to adapt quickly enough. I like the image of dinosaurs standing in a field watching fire pour from the sky.

Stephen Arnold, December 7, 2009

I have to confess that I wrote this because Tess licked my face. I suppose I need to report this to the Department of Health & Human Services. This outfit is responsible for unsanitary inducements from a rescue boxer with a physical deformity. I quite like Tess’s payoffs, however.

Surprises in Search: Holiday Gifts for Pundits

2009-12-05T22:19:39Z

The first week of December 2009 delivered a rich feast to mavens, pundits, poobahs, and azure chip consultants. Kwanza and Chanukah arrived early.

What’s transpired in the last week or so?

First, Google jumped into eCommerce search, hired Endeca’s chief technology officer, and made it clear that it wanted to suck in some cash from this lucrative sector of the information retrieval market. I know that most of this eCommerce excitement happened over a period of weeks, but the significance of this tactical move is significant. Endeca which has been working through its cash injections from Intel and SAP now has a real fight on its hands. The company has an RV full of bright MBAs, and it will need these folks to convert a Google thrust into new revenue. Fun holiday and New Year’s Day ahead for the Endeca folks. My “cutting Endeca from the herd” story evoked a gust of push back when I voiced my opinion. I stand by my argument in “Google Cuts Endeca from the Search Herd”.

Second, Microsoft and Yahoo finally got married. Well, maybe the right word is decided to implement their version of the Bernard Slade play “Same Time, Next Year.” The two outfits are going to team up to challenge Google. If you are not on top of this unusual sales / search deal, you may find the Macworld write up a useful summary. Read “Microsoft, Yahoo Finalize Search Deal.”

Third, Oracle, a notoriously aggressive outfit, attacked Mark Logic in a white paper filled with interesting assertions. I wrote about this on December 3, 2009. If direct attacks escalate, 2010 will be a contentious year. The notion that a multi billion dollar database company is threatened by a next generation XML data management system says more about the fear that traditional companies have for smaller firms with better technology.

Finally, the odd yet interesting machinations of traditional publishing companies have signaled a perceived urgency in adapting to the digital information environment. There’s the dust up between Google and News Corp. over indexing. I noted the San Francisco Chronicle’s take on one facet of this battle in “Google’s Schmidt Strikes Back at Murdoch.” More interesting is the push by traditional print publishers to create an online newsstand, sort of a modern day version of the old “Magazine Rack” idea from 15 years ago. You can read about this “old wine in new bottles” approach in “Publishers Alliance to Create iTunes for Print Media as Tablet Rumours Continue to Build”. The problem with this idea is that it is aimed at a generation of readers. The demographics argue that readers will be a small percentage of the consumer population which may not bode well for dreams of new revenue for traditional publishers.

I am looking forward to the discussion of these holiday season topics.

Stephen Arnold, December 6, 2009

Oyez, oyez, I wish to disclose to Court Services and Offender Supervision Agency for the District of Columbia that I was not paid to write this round up of holiday delights.

IBM Asserts that DB2 Is Better, Faster, and Cheaper

2009-12-05T22:18:49Z

I thought that in the old joke about “better, faster, and cheaper”, I could only pick two attributes. The idea, of course, is that the three attributes don’t live in the same Venn diagram like lovers in a trailer court in rural Kentucky.

Take a look at “IBM DB2 Gains New Enthusiasts in Search of Better Performance at Lower Cost” on the IBM.com Web site. Note this subhead: “More than 100 SAP Clients Have Switched their Database Software to DB2 in Last Six Months”.

The idea is that SAP customers whom I assume * were * using Microsoft SQL Server jumped to IBM’s DB2. Now SAP has its hands full with grousing customers, revenue decreases, and competitive pressure. The guts of the news release is that when the “old dog” is replaced with a “new IBM DB2 breed”, staff costs go down by 35 percent and “efficiency” (whatever that word means) shoots up by 65 percent.

The key paragraph for me was this one:

Since 2003, IBM DB2 on POWER Systems has demonstrated consistent performance leadership. DB2 uniquely holds the number one position in the online transaction processing (OLTP) (1), warehouse (2), and SAP standard application benchmarks. Recent results on the two-tier SAP Sales and Distribution (SD) Standard Application Benchmark showed that IBM DB2 running on a POWER 550 System delivered high performance of 3,752 SAP SD Benchmark users (3).

My hunch is that a price war among major RDBMS vendors is brewing. I don’t think the performance differences among these long-in-the-tooth RDBMS systems is achievable without some serious hardware. The next generation data management systems are likely to become more attractive. Big Blue and its ilk cannot afford long term price wars. Billing for services to make up the license losses won’t do the trick either. In short, spin before the storm.

Stephen Arnold, December 6, 2009

Okay, I confess. I was not paid to write this opinion. I bet the person who Twitter the link to me, a certain IBMDB2, was paid, however. I wonder if he will report to the Court of Appeals for the Armed Forces as I have? Probably not. I seem to be one of a small percentage of Web log authors who report who pays me and who does not to anyone. A lone goose again!

London Online: The Missing Trends

2009-12-05T19:31:55Z

The endnote at the International Online Conference succeeded in getting insightful comments from the panelists and eliciting probing questions from the audience. The downside was that the 90 minute session covered four of the 10 trends advertised in the program. The four trends discussed in the endnote were rising Google pressure, more use of XML, a surge in rich media for core information exchange, and more security safety nets with increased user surveillance likely.

In response to several emails from attendees, here are the missing six trends:

Trend 5: Libraries will be under increasing budget pressure. As a result, interest in lower-cost, cloud-based solutions will rise sharply in 2010. One consequence will more financial woes for library vendors, including commercial database producers.

Trend 6: More demands for timely data. Although not real-time indexing and content delivery, the newer services will strive to reduce latency (staleness) of information available to users in an organization.

Trend 7: Mobile search will become more important. The impact on the length of certain types of textual information will be significant. Those without fast network connections will be unable to access the rich media that will become a larger percentage of the information on offer.

Trend 8: Even if the economic climate improves in 2010, there will be increasing financial pressure on information, search, and content processing companies. Content management and enterprise search vendors will be particularly vulnerable. Neither CMS nor search can “explain” precisely their benefits so marketing, not technical excellence will mean the difference between survival and a buy out or extinction.

Trend 9: Open source will gain traction. Traditional vendors will have to deal with the financial and technical payoffs open source offers. In some organizations, open source will become an acceptable alternative to certain software systems. At the same time, open source vendors will monetize their services. Confusion and contention will increase.

Trend 10: Regulation will become more oppressive. In 2010, the Wild West of the Internet will be brought under the control of the authorities.

Have a trend to add? Use the comments feature of this Web log.

Stephen Arnold, December 6, 2009

Yep, I was paid to be at the Incisive show by Incisive. Nope, I was not paid to write my view of the trends in 2010. Deal with it.

Google Programming Preferences, the Python Ultimatum

2009-12-05T02:48:08Z

A thread on Google Groups’ Unladen-Swallow tackled the question of what programming languages are encouraged or discouraged at Google. At the conference in London on December 2, 2009, I heard that Microsoft wanted developers to use Visual Studio.net, yet Windows 7 and other Windows products were anchored in Assembly language and C, often undocumented Assembly language and C. Whether true or not, the resulting conversation triggered on person to mention the Unladen Swallow thread.

I hunted down the group and noticed several interesting points.

First, one group participant asserted that Python is not as zippy as Java or C++.

Second, Unladen Swallow included this passage:

Unladen Swallow aims to shift the balancing points in that tradeoff to make it possible to use Python in more places where it would currently be unsuitable, but it’s not going to be a panacea. Python will still be slower than C and Java, use more memory and have inferior threading until someone decides to invest resources into Python comparable to what, say, Sun has invested in their JVM. I hope a focus on Python performance by the developers will start a snowballing effect: more companies are interested, more resources can be devoted, more grad students will work on Python (and actually commit their work), etc.

Third, on the issue of JavaScript, there were some comments. Among them was this one:

I don’t think it’s possible to make an implementation like CPython as fast as an engine like V8 or SquirrelFish Extreme that was designed to be fast above all else.

Fourth, on the subject of Jython:

Another one of unladen-swallow’s major goals is to maintain source-level compatibility with C extension modules, which Google uses a lot of. Using Jython would require moving that infrastructure from SWIG to JNI, which would be a huge pain in the butt.

To which, this comment was added:

Jython has a similarly-small number of paid full-time developers (and Frank W just left Sun, as you may have seen). To date, IronPython and Jython have had to dedicate a significant portion of their engineering resources to achieving compatibility with CPython, with relatively little time left over for performance optimization. It remains to be seen how these factors will be balanced as Jython, IronPython, PyPy, etc move to support Python 3.

Finally the comment I found most interesting:

One area where I believe Python (and other dynamic languages) do particularly well is projects with lots of developers working on lots of little subsystems, all contributing to the same codebase or the same released binary. In a language like C or C++, the more developers you add, the more fragile your binary becomes: it only takes one segfault to kill a running binary (and hence lose those pending requests), and the probability of introducing that segfault goes up with the number of developers/subsystems/integration points/etc. Dynamic languages, on the other hand, are much easier to sandbox in this regard. If you want to isolate failures in one particular component in your Python system, you can just throw a try/except around it and you’re basically good to go. That kind of agility and flexibility is an important advantage, and it’s one of the “other merits” I was alluding to above. Python may not be a good fit for millisecond-critical systems, but it
does have advantages; we wouldn’t use it if it didn’t.

What is clear to the addled goose is that this discussion of performance at a fine-grain level is part of Google’s now public effort to make the Web speedier. The addled goose wants to point out that with each “make the Web better”, Google takes a baby step closer to becoming the Web.

This is a big idea but it grows from fine grained engineering emphasis on performance details. Now about that Office 10 beta load time?

For criticism of some Google practices, navigate to Google Releases its JavaScript Closure Tools. For a critical discussion of the Google Go language, check The Most Intriguing Concept In Google’s Go Language. More information about Google programming appears in my Google trilogy.

Stephen Arnold, November 2009

Again I must report to the General Services Administration that I was not paid to write about a performance discussion of Python variants at the Googzilla facilities. Speedy “services” remain the exclusive domain of the General Survives Administration, so Google has a long way to go to achieve GSA velocity.