Aweekstweets Sept 20-27 2009 @jameskobielus
BTW, has anybody noticed the irony of the so-called "father of the Internet" being somebody named "Cerf"?4 minutes ago from TweetDeck
JK2—Surfing the Internet is still an apt metaphor. New waves of innovation just keep swelling this ocean of information and experience. It’s all we can do to stay on the board.
Scanned & posted to my Facebook key docs from 1991 attesting to Gore's pivotal role in commercializing the Internet. Setting record straightabout 8 hours ago from TweetDeck
JK2—Gore was one key catalyst in the development of the first truly new mass medium since television. He never claimed to be the “father of the Internet.” In fact, Vint Cerf—another key catalyst--never claimed that (admittedly sexist, paternalistic, simplistic, GreatMan-istic) role either. It’s simply a matter of Gore’s political enemies exaggerating what he claimed in order to tarnish his legacy and keep him from gaining the presidency.
"Cerf: Turning off pieces of the 'Net 'not sensible'" (http://bit.ly/15QDP3): JK--Bogus issue. No one's arguing POTUS needs to be Net-fuhrerabout 9 hours ago from TweetDeck
JK2—With reference to the proposed Snowe-Rockefeller legislation, Cerf sets the record straight in the linked article: “You understand the motivation behind it. It's people concerned about the dependency we have on the Net and the potential for that dependency to be very brittle. But the idea that the president would turn off pieces of the Internet is not sensible. On the other hand, focusing attention on the need to make the network more robust, more reliable, more resistant to various forms of either attack or infection is a really good thing to be attentive to. .... I don't think the president should try to turn off the Internet, and I don't think that he has any interest in trying to do that.” Heck, Bush Jr couldn’t “turn off” the civil aviation system in this country for long after 9/11? And what, exactly, would one “turn off” in response to a cyber-warfare attack? And isn’t “turning off,” or blocking, pieces of the Internet the preferred solution of the Chinese communists?
Anyone up for a Delete-a-Thon? How fast can you manually detect/purge spam from various accounts? It's everybody's everyday obstacle course.about 10 hours ago from TweetDeck
JK2—With a good e-mail client, a keyboard and mouse, a fast connection, and a net-side spamfiltering service, I can clear out a hundreds of spam in perhaps 30 seconds—with very high accuracy--from both my inbox, client-side bulk-mail folder, and online suspected mail folder. I have several e-mail inboxes, so we’re still talking a couple minutes of my time, at least, each and every morning.
RT @smarthart: "MRotenberg discusses modern day privacy." http://bit.ly/hW2vc: JK--It's power to make oneself less of a sitting-duck targetabout 10 hours ago from TweetDeck
JK2—Much of this is a “cows long since left the barn” discussion. Says Rotenberg: “Is the government going to use our data as it is supposed to or is it going to spy on us? Does the bank have good security or do we have to worry about breaches? If I give an email address to a cell phone company, am I going to get spammed? And is that quiz that just told me which European capital city I'm most like really trying to figure out who my friends are?” It’s also about the power to monetize your online self in all its informational splendor, including your blogs, tweets, and facebook status updates. Says Rotenberg: “This is modern-day privacy, about digital identity, the control of personal information, and the brewing battle between what we post and its commercial value.” Every last scrap of what we choose to share online will in fact be scooped up by the growing silent army of “listening engines” that leverage text analytics technologies: natural language processing, entity extraction, sentiment extraction, social network analysis, and the like. Some of these engines will correlate your profile, behaviors, and transactions based on your unique identity. Others will roll it all up into broad demographics and psychographics, to fine tune their strategy of reaching you and people like you. Try as you may to stop this silent army from targeting you with every last scrap of intelligence that you choose to share (or don’t realize you’re choosing to share, or others are digging up on you from disparate sources) you won’t succeed. Best to prepare your defenses. More powerful spam filtering, for example.
RT @chaskielt: R @kalido: Had in-depth conversation about Kalido KONA w/ @jameskobielus." JK--Kalido partnering with Netezza and QlikView.2:27 PM Sep 25th from TweetDeck
JK2—Forthcoming (not yet GA) solution appliance for master data management, with out-of-box information models (initially) for pharmaceutical sales and marketing analytics and insurance sales and distribution analytics. I haven’t yet come across a customer looking for a shrink-wrapped appliance for MDM. Usually, DW appliances are for query acceleration in a mart layer in front of an existing EDW hub. But give Kalido credit for pioneering a potential new market segment. KONA will become available in roughly the same timeframe as IBM’s pre-announced Smart Analytics System solution appliances (of which it has not yet announced specific vertical and horizontal applications and markets that will be addressed).
Just responded to Forrester customer email inquiry on BI-in-cloud. Reviewing research, I'd forgotten how many vendors in this immature space9:22 AM Sep 25th from TweetDeck
JK2--There is a growing range of vendors that offer DW in a public and/or private cloud environment, including Vertica, Kognitio, Aster, Greenplum, Microsoft, AppNexus, FathomDB, and 1010data. There is another group of vendors that specialize in offering the front-end reporting, query, dashboarding, OLAP, and other access, presentation, interaction, and delivery features of BI as cloud/hosted/SaaS subscription service. This list includes Bi3, Birst, BlinkLogic, FortiusOne, Good Data, InetSoft, JasperSoft, Oco, PivotLink, and QlikView, some, but not all, of which incorporate a hosted DW into their hosted BI offerings (all of which are managed from their own data centers, not public clouds). Furthermore, many leading BI vendors—including Actuate, IBM Cognos, Oracle, Panorama, SAP Business Objects, TIBCO, and QlikView—offer SaaS subscription access to their solutions (but none of them yet hosts these services in Amazon EC2 or any other public cloud).
Handled reporter inquiry on Oracle Exadata v2 last night. Key new feature: flash cache. Paves way to eventual all-SSD, all-in-memory DW/OLTP6:57 AM Sep 25th from TweetDeck
JK2-- Performance-wise, one of the most significant enhancements in the Exadata v2 machines is the cache-oriented database-acceleration technology (5 TB smart cache per machine), paving the architectural road toward an eventual phase-out of physical disk in favor of an all solid-state-drive (SSD) storage architecture for order-of-magnitude I/O bandwidth expansion. The new Sun-based machines, leveraging Exadata Smart Flash Cache (based on Sun FlashFire memory cards), offer significant improvement in Oracle Database 11g R2 query performance, and also in write-intensive online transaction processing (OLTP) applications.
Edited Wikipedia entry on Web data services (http://bit.ly/x2Iqs).7:28 AM Sep 24th from TweetDeck
JK2—A couple of days later, my revs stand unmolested. Here’s what it says up there as of this moment:
“Web data services [1]refers to service-oriented architecture (SOA) applied to data sourced from the World Wide Web and the Internet as a whole. Web data services enable maximal mashup, reuse, and sharing of structured data (such as relational tables), semi-structured information (such as Extensible Markup Language (XML) documents), and unstructured information (such as RSS feeds, content from web applications, commercial data from online business sources). In a Web data services environment, applications may subscribe to and consume information, provide and publish information for others to consume, or both. Applications that can serve as a consumer/subscriber and/or provider/publisher of Web data services include mobile computing, web portals, enterprise portals, online business software, social media, and social networks.[2]. Web data services may support business-to-consumer and business-to-business information-sharing requirements. Increasingly, enteprises are including web data services in their SOA implementations, as they integrate mashup-style user-driven information sharing into business intelligence, business process management, predictive analytics, content management, and other applications, according to industry analysts.[3] To speed development of web data services, enterprises can deploy technologies that ease discovery, extraction, movement, transformation, cleansing, normalization, joining, consolidation, access, and presentation of disparate information types from diverse internal sources (such as data warehouses and CRM systems) and external sources (such as commercial market data aggregators).[4] [5] Web data services build on industry-standard protocols, interfaces, formats, and integration patterns, such as those used for SOA, Web 2.0, Web-Oriented Architecture, and Representational State Transfer. In addition to operating over the public Internet, Web data services may run solely within corporate intranets, or across B2B supply chains, or even span hosted software as a service (SaaS) or cloud computing environments.
Finding the mobile carrier industry the chief proving ground for comprehensive in-db, in-cloud, and in-stream analytics.3:10 PM Sep 23rd from TweetDeck
JK2—The key apps for all these are customer churn analysis, customer up-sell/cross-sell prospecting, customer experience analysis, and fraud prevention/detection. Predictive models—embedded in the data warehouses, cloud computing environments, and real-time complex event processing streams upon which mobile carriers run their operations—enable carriers to maintain proactive 24x7 surveillance, look-ahead, and optimization of their environments on every level through fearsomely smart rules engines.
"Teradata: 'We were the original appliance vendor'" (http://bit.ly/1L2vTI): JK--Remember when DW vendors avoided the A-word like plague?1:22 PM Sep 23rd from TweetDeck
JK2—Nowadays, if a DW appliance vendor avoids the word “appliance,” it shows they have no clear sense for what customers are screaming for: one-stop shop for a complete, pre-integrated, pre-optimization solution portfolio on a hyper-fast hardware/software platform.
Expect DW appl vendors to push several envelopes in pitches: sub-$20K/TB affordability , super-100X query accel, & MR in-db analytics flex.11:25 AM Sep 23rd from TweetDeck
JK2—Maximum affordability, scalability, and flexibility. The MapReduce in-database analytics must be comprehensive enough to encompass pushdown of all functions—ETL, query, regression, scoring, etc.—into a massively parallel, distributed, virtual DW fabric.
Awaiting further price reductions among DW appl vendors. Expect sub-$20K/TB deal vol to grow, for core commodity DW, not incl add-ons & p.s.11:17 AM Sep 23rd from TweetDeck
JK2—DW deals for under $20,000 per terabyte (compressed, user data) for the platform (CPU, storage, I/O, query tools, loading tools) maintaining the core production tables. The sub-$20K/TB deal volume will consist primarily of deals in which that lowball pricing is accompanied by long-term customer subscriptions for maintenance/support, plus add-ons such as BI, ETL, DQ, and MDM, plus a host of professional services to put it all together, customize it, and optimize it all anew.
RT @dmscott @gmblogs #gm "Social media democratizes word of mouth." http://bit.ly/1y1WKQ : JK--Becomes word cloud of unseen flying fingers.8:39 AM Sep 23rd from TweetDeck
JK2—Most of which will feel unreal. Much of which will in fact will be auto-generated by spambots targeting suckers. Much of which will purport to come from reputable sources, or people who’ve “seen it, done it, used it” and whose testimonials must be believed. Caveat caveat caveat!
Analytics vendors need to maintain healthy "in-the-lab"/"in-customer-hands" balance on their techs. Too lab-weighted = PhDs gone wild!7:39 AM Sep 23rd from TweetDeck
JK2—I love the stuff in the lab. I could spend all my days and nights in the lab. I’m an analyst who sees countless me-too products. The stuff in the lab is usually the antidote to boredom—the seriously new stuff—for jaded old pros such as me. Then again, the people in the lab aren’t talking to customers and trying to develop and sell products. Caveat caveat caveat!
In business analytics, a "full-stack vendor" now refers to one that offers comprehensive hardware, software, and services portfolio.7:36 AM Sep 23rd from TweetDeck
JK2—“Full stack” doesn’t mean you should acquire everything or anything from that vendor. It just means their product/service catalog should guide your checklist of what, if you had all the budget in the world, you would in theory need for a feature-complete deployment of whatever.
Price wars! One DW appl vendor touting so-called "Cash4Clunkers" (3-yr sw subscr: #1 free, #2&3 cut-rate) incentive to migrate from rival.10:18 PM Sep 22nd from TweetDeck
JK2—I love when vendors disparage each other in the context of too-hip sales/marketing campaigns. Already, “Cash For Clunkers” feels a bit past its expiration date as a buzzphrase in the zeitgeist. But the DW appliance price war continues unabated.
RT @atanubasu: "Social media analysis: a new way of listening" http://bit.ly/buNdb: JK--Or new way of pigeonholing, with stats on your side.10:03 PM Sep 22nd from TweetDeck
JK2—Social media analysis drives clustering and segmentation, which in turn drive target marketing. From the recipient’s end of the target-marketing equation, it all begins to blur into spam and other invitations/solicitations that feel like they’re pitched at somebody else, not actually or specifically you. Trying to make your feel like you should be part of some “social network” that you don’t feel like you belong to, or which doesn’t add meaning to your personal experience.
RT @jackshafer: "Clay Shirky on investigative journalism” http://bit.ly/P5qhn: JK--Who'll replace reporters as watchdogs as newspapers fold?9:51 PM Sep 22nd from TweetDeck
JK2—Answer: analysts. Further context on that answer: that’s what the best reporters have always been, and most of the best are now bloggers.
"Netflix Awards Prize" (http://bit.ly/17N0um): jk--3-yr $1M contest to tweak movie-recommendation engine predictive model. Razor-thin finish9:46 PM Sep 22nd from TweetDeck
JK2—Is Netflix going to hire these people to score and evolve the model? It will decay, of course, and soon.
Most Hadoop being implemented in premises "clouds" (i.e., distributed, virtual, complex info/file stores), not public clouds.9:31 PM Sep 22nd from TweetDeck
JK2—Hadoop devotees bristle at being associated with in-database analytics alone.
RT @SethGrimes " DBMSs before electronic computers...edge-notched cards" JK--Wha? Any stackable non-volatile inscribed surfaces a database?5:44 PM Sep 22nd from TweetDeck
JK2—Let’s not imagine “databases”—a late 20th century innovation--as the pinnacle of human informational evolution. That is uncomfortably close to the “everything that’s ever going to be invented has already been invented” foolishness.
RT @nenshad: "Twitterese is "domain-specific language" in CS speak." JK--Yep. DSL for chatter-intensive attention-poor trivia-steeped domain5:41 PM Sep 22nd from TweetDeck
JK2—Let’s not imagine Twitter, in its current state, as the pinnacle of micro-expression on the Internet. Telegraphy gave way to paging, instant messaging, and SMS, then Twitter, and others will surely come. None of them created a new language—just a new constraint into which the old language mapped its complexity and stripped away distracting nuance. Essentially, each new micro-expression medium creates a new pidgin English (look up the technical linguistic definition of “pidgin”).
"The Perils of Cameras in Mobile Devices" (http://bit.ly/YtrKh): jk--Yeah. But also peril to collective memory from not having camera handy8:14 AM Sep 22nd from TweetDeck
JK2—Just keep snapping those camera photos. Future generations will thank you for “being there.”
What exactly is a "Luddite" anyway? I try not to complicate my tech life anymore than necessary. Only on my own terms. And when good & ready7:44 AM Sep 22nd from TweetDeck
JK2—For example, these aweekstweets.
Liking the new TweetDeck v0.30.5--the Facebook integration is much better, richer.3:41 PM Sep 21st from TweetDeck
JK2—But the new TweetDeck has its issues: the auto-URL-shortener is unreliable, and the Facebook link tends to de-select for no apparent reason.
RT @computerworld: "Dell agrees to buy Perot Systems for $3.9B" JK--HW a commodity, prof svcs not. Both Perot companies bought by HW vdrs.8:04 AM Sep 21st from TweetDeck
JK2—Sure am glad the little demagogue didn’t become president. Whatever became of the sucking sound he said NAFTA passage would cause as US jobs went down to Mexico? Last time I checked, Mexicans still flocking to the US, even in recessionary times. Good for them. Glad to have them here. And it’s good for them to know that US companies are not averse to investing south of the border, should they decide to go back home. My crystal ball tells me Mexico will become one of the world’s most powerful economies in this century.
