Share This

Share This

Download the episode directly right here, subscribe to the feed in iTunes or other podcatcher to have episodes downloaded automatically, or just click play below to listen to it right here:

John and I caught up earlier in the week. Despite it being a short time between this episode and the last, we found plenty to talk about:

• Man! It’s hotinhur!
• John’s Cloud Week videos - John Adams from Twitter, Chris Wensel on Cascading, ControlTier.
• The “science” meme at Velocity - R, Mathmatica, and doing stats. Coté also remembers that Sara Dornsife now works for the R company, and the Zed Shaw talk that included R. I ask John what you’d use this stat stuff for, obviously charting historic data, but also doing predicting.
• Chef vs. Puppet - also cfengine is mentioned in this context. What other “alliances” does OpsCode have around Chef? EngineYard, (maybe, if John remembers) RightScale. Also, they got $2.5M in funding (we mistakenly remember $5M).
• Eucalyptus - talking through their layered up, (seemingly) “swappable” cloud platform and how they’re thinking of making money off it.
• IBM CloudBurst demo (this one, or this one, or another one? - how was that? John likes the Service Management hook-in, self-service portal.
• Is “private cloud” slowing down “cloud computing.” And getting at traditional IT shops “hav[ing] that ability to, kind of, ‘API’ themselves.”
• Are SLAs and KPIs calling out for “better” operational stuff in IT like private cloud?
• This gets us into webappVM - deep instrumentation; “webappVM is building a self-monitoring application cloud for web applications, bringing enterprise-level application management capabilities to public and private clouds” from their site.
• FiveRuns Dash - generic-ish, Web 2.0 style metrics. This gets us onto FiveRuns memories. Also, former FiveRuns CEO OT is now at Zenoss as their CMO.
• Google Ops Head goes after blood at Structure 09, .Net at MySpace
• Kuleto’s review
• John [was] helping run CloudCampColumbus this week. He’s trying to get some Tivoli folk out there.
• Coté interested in pictures and anecdotes of Detroit. [We'll see if John snatches some.]
• I mention that I’m thinking of strumming up interest in barcampESM II (see the first for reference) - is there enough interest for that kind of thing this year? Also, I’m lazy.

Disclosure: Reductive Labs (Puppet), IBM, and Zenoss are client.

Share This

In this interview from RSC 2009, I talk with IBM’s Neeraj Chandra, who you may recall from two previous interviews at RSDC 2008 and Innovation 2008, last year.

We start out talking about what exactly a “Smart Product” is and how Rational fits into the overall IBM “Smart Planet” vision. Here, the discussion gets into one of the recent Rational tenants: businesses should not only be looking to software for differentiation and value, but are indeed forced to.

The question then, is how IBM helps companies do this: the goals are, of course, desirable, but the devil is always in the details. Part of the story here is the need to bring more discipline to the software creation process as it raises is criticality to the business.

While the IT-side of the equation has to change, there’s also much needed from the business side. We discuss how the business-side needs to change and adapt to these scenarios as well. Neeraj points out that much of this change is enabled by upping the collaborative aspects in the overall Rational portfolio - enabled, of course, by the Jazz platform.

We get back to to how business strategy and objectives map down to IT and the development of software - I ask Nerraj to go over how bodies of practice like Rational’s MCIF are used to map between the two sets of objectives.

Disclosure: IBM is a client and sponsored these videos.

Share This

Yesterday I asked the “my tweeps” if they’d like to see notes on all those briefings. There wasn’t a “no” answer, but some people said they’d prefer to see just the good ones. I loath making content-producing promises - because I always break them - but so far I have three “brief notes” for those interested:

• MindTouch Aims Take Over Your Intranet
• Apache Pivot
• LongJump, Pass in a Box

While I don’t plan on writing up every single briefing I get - some are confidential, others just boring, and then everything in-between - it is (professionally) nice to do so. I’ll see if I can set aside some time each week, as I did this week, to write them up. If you like them, leave a note so I know to keep it up.

Share This

LongJump is one of these outfits who’ve I now spoken with a couple times, including earlier this week with CEO Pankaj Malviya. Put in brutally simplistic terms, the provide a RAD stack of software that helps developers create web applications backed by databases. A hugh amount of business applications - and consumer - are really just that: a web interface on-top of searching and changing a database, so that’s no poor niche to be in.

From RAD to PaaS

Back in February, LongJump started providing their platform to ISVs packaged to be used as a private PaaS, where “private” means “run on your own,” just as “private cloud” does. That is, these ISVs would use the LongJump platform to develop and then sell SaaS applications. This week, their CEO Pankaj Malviya updated me on the progress of this offering. We walked through a nice demo, including new features like “workflows.” “Workflow” in this context usually means a process involving multiple people who have all go through a cycle along the lines of: the creation of some task at hand, assign someone to do something with the task, get someone else to “approve” that the task is done. Throw in a “customer,” and you’ve got “BPM” ;>

Since February 2009, they’ve had 5 ISVs sign up, each creating of delivering SaaSes with LongJump in very narrow industries like electronic equipment maintenance, HR, and drug trial management.

PaaS over the ALM

One of the more interesting statements from Pankaj was that development shops were finding the lack of traditional (to use my modifier) ALM in PaaS development annoying. Folks like Force.com - and the wider “cloud” world” - the conceit goes, don’t provide enough hooks to fit into the software development process that people are used to. The thinking is that there’s a cultural mis-mapping between how people want to develop software for the cloud and how the cloud wants (if only by omission of other options) people to develop software. Bringing up the specter of ALM-weakness is a good ploy and segmenter: if you know what ALM means, you probably like it, and if you don’t know what ALM means, you’re probably not worth selling to. It’s like command-line vs. “dashboards,” in that respect.

Share This

While Adobe, Microsoft, and Google/HTML5/Mozilla are battling it out to capture the UI-space of the 2010’s (how are you supposed to say “2010’s”? “The teens”?), there are a handful of other efforts to provide a post-Ajax UI layer. The Apache Pivot project, now in incubation, is one I came across recently. Essentially, the Pivot team - mostly from VMWare - have been working on Java-based GUI UI from the ground-up.

Pivot Overview

Rather than use existing (and much despised by all - see JavaFX) Java GUI toolkits like AWT or Swing, they specify their own UI markup language (WTKS), which then sits directly on-top of Java2D and other GUI widgetry to render UIs. Thrown into this mix are concerns for web-based networking (like running in the browser’s Java plugin and providing server side components for data access), and you essentially have a framework for what we’d call a “Rich Internet Application” (RIA) now-a-days, minus the video. I’m always more into the UI- rather than the video-side of RIAs, so that’s really no issue for me.

Architecture, etc.

Pivot is divided into a few broad buckets:

• core - non-UI classes for utils and such
• web - platform support for server communication; REST
• WTK - windowing tool-kit, component definitions and then another library that provides the skin (like Swing)
• charting package - plugin to JFreeChart

For more technical detail, check out the Apache Pivot posts from Gregg Brown over at InsideRIA.com.

The RIA Market

Being at the ASF, Pivot is going through the incubation process, meaning it’s getting vetted and otherwise “put into shape” to be a full fledged Apache project. I wouldn’t expect Pivot to be a major market competitor for the likes of the Flash Platform, Silverlight, or even Ajax anytime too soon.

That takes a dedicated commercial and marketing effort which the ASF really doesn’t care to get involved in itself, and I don’t get the sense that the Pivot dudes will extract a Covalent or Cloudera out of the project anytime soon. That said, as the spread of Eclipse’s UI layer has shown over the years, there’s a deep yearning in the Java community for a better GUI solution. Even Sun’s conceit with JavaFX is to deliver on that desire. Closed source GUI frameworks have a tough time at it now-a-days, where-as open source ones by virtue of being free and open, potentially have an easier time to dig into the minds of Java developers.

Disclosure: Adobe, Microsoft, and Sun are clients.

Share This

Previously, MindTouch primarily sold a framework for building up collaborative business applications. Now they’re putting together finished applications targeted at either specific uses - like groupware intranets - or industry verticals. Their first - out of three - product here in this space in MindTouch Collaborative Intranet.

This marketing position is subtle, but the point is to go after the application space rather than the foundation/framework/developer space. Their CEO Aaron Fulkerson has written several times about the difficulty of specifying and, thus, carving out a “collaborative networking” category - really, much of the idea of what all this “Enterprise 2.0″ stuff is fussing about. A table in one of Aaron’s recent pieces draws out what makes something a business application vs. a social applications (he has his own terms for both):

MindTouch is and can be used in all sorts of web applications contexts. It’s very common to see it in shoot-outs with online community engines (this from MindTouch’s early focus on wikis), but as this new positioning shows, it’s geared up to be a SharePoint Killer as well. More than just the specific product of SharePoint itself - which rival vendors will all but describe as intranet kudzu - the whole concept of “that software groups of people use behind the firewall to collaborate” has fallen under the SharePoint concept-space: very analogously to “Office” meaning more than just Microsoft Office in the word processor, presentation, and spreadsheet category.

“Portals,” as a category also fill this intranet category, and IBM Lotus among others sells into this box as well. I rarely hear people raving about their “portals,” and it’s clear from the rate at which vendors pull features from the public web that, in most cases, the combined collaborative services that the public web brings are more pleasing that what behind-the-firewall collaborative technology does for employees. Part of this has to do with cost - public web sites are free, for the most part, compared to business apps which are not - lack of legacy support and cultural differences in use. All that said, each time I see someone like MindTouch trying to finally make intranets more like the internet with both culture and technology change, I’m hopeful we’ll finally crack it.

Disclosure: MindTouch, IBM, and Microsoft are clients.

Share This

In this interview from RSC 2009, I talk with IBM’s Mike O’Rourke, who you may remember from last year’s RSDC series. We discuss this year’s RSC focus on reporting, metrics, dashboards, etc. vs. “low-level” developer tools. As Mike says, this solves a problem where people are “tracking activity in a tool, but not across a whole life-cycle [of software].” Embodying this is the the Insight tool that tracks a requirement across the life-cycle of development, QA, deployment, and so on.

We also discuss how these tools fit in with best practices and establishes methodologies of companies. How do you get a development organization to the point of maturity where they can even start to track how well they’re doing? Mike spends discusses how he does this across the projects he overseas in the Rational portfolio, and maps it back to the generalized case. This also leads into a discussion of the recently released MCIF and how the process therein is hooked into the tooling and reporting in development. I then ask Mike about the team and organization size that Insight, MCIF, and other products work well with.

Shifting gears slightly, I ask Mike to tell us about the past year’s efforts to incorporate the Telelogic tools into the Rational portfolio. Part of this is the ongoing effort to have Rational products built on-top of Jazz, so Mike goes over the road-map for those transitions.

We wrap up with a discussion of the issues that Rational has encountered in moving some of their products to the cloud. I ask Mike for some lessons learned that might be helpful to others doing the same. One interesting practice Mike mentions is getting previews out there to enable richer customer/development feedback on the product, rather than waiting for the 12-18 month big-bang feedback cycle.

Disclosure: IBM is a client and sponsored these videos.

Share This

Share This