Live blogging web 2. 0 expo – ibm and performance

See table of contents for full list of web 2.0 expo posts

Unlike Blackberry, IBM’s looks more advertising thsn Blackberry so they can share a blog post with others. I’m not writing a commerical. The performance part later in this blog post was great. Also blogged about payments in here (didn’t make title because I didn’t think i would be back in time for this session). I will write more about the keynotes in the afternoon.

IBMs application transformation, modernization and revitalization to a web 2.0 experience

  • Problem: what to do about the mainframe
  • Solution: pay IBM. Sounds somewhat magical.
  • I was expecting more because the last time I saw an IBM session at this event, it was about social networking and showed how Connections was used at IBM. It was a commercial, but didn’t feel like one.
  • Did demo of turning a green screen app into a web and mobile version. Cool that the screen was originally copyright before i was born. Also thinking back to when the public library card catalog was a green screen; didn’t know what a terminal was then.
  • Concept interesting – using rules and transformations to call mulitple mainframe screens. Of course, the mainframe still exists this lets use it thru web.
  • However, not clear on what this has todo with web 2.0. An IBMer said javascript sorting is a web 2.0 feature. Not by my definition.

Why you have less than a second to deliver exceptional performance – dynatrace/compuware

  • Book: “Designing and Engineering Time” – how people perceive time
  • Book: high performance web sites
  • Book: high performance javascript
  • Instantaneous less than .2 second – like clicking button/pull down
  • Immediate less thsn 1 second – like scrolling or paging. Because think info is already there
  • Continuous less than 4 seconds – like when asking a person a question because think time. We expect request to system to take time as well. 2 seconds for something simple like home page. 4 seconds for query


  • Can stay focused on a task 7-10 seconds. By then our attention moves to something else like e-mail. Shouldn’t take 7 seconds but an upper bound. 7 seconds was early recommendation as upper bound and we got used to it training or patience to that time.
  • In last three years, people got 50% less happy woth 4 second response time
  • We cannot perceive a 20% time difference so need oess than 1.6 seconds to be perceived as being exceptionally fast. And that time inckudes network, dns lookup, rendering, etc
  • Even if bandwidth high, high latency (travel time) still affects perceived performance. Broadband is 300ms latency. A lot when shooting for 1600 ms end to end.
  • 200 kb at 1.5 Mb/s takes a second
  • Client rendering typically takes .3 seconds.
  • Firefox and chrome render much faster than ie or safari 4 (safari 5 only slighly higher)
  • – compare your site to ohers

From Intent to expression @bsaren from Litle on payments

  • He had a slide on Occupy Wall Street. Nice to see the slides are current. The point being things are changing in finance.
  • Lines bluring between consumer and professional services like vimeo.
  • Bitcoin is virtual currency. People are hording it like it is gold. [sounds familiar; I think I read about this in the paper]
  • Privacy discussion needs to happen. Thinks will be differentiator going forward. [profiling by my type of credit card feels weird]
  • payments intelligence – recurring payments vs person with prepaid card with money for one month
  • it was interesting but more business side than tech side so didn’t capture much content

Secret Sauce from Yottaa

  • Now that everyone has a website, user experience becomes the differentiator including performance
  • SEO – skow sights rank lower on google=
  • Impressive stats on financial impact of just one second slower load time
  • Facebook and twitter widgets make site slower
  • Client/browser side often adds more time than content delivery

And now on to the day 1 keynotes

Leave a Reply

Your email address will not be published.