Disqus - Latest Comments for gengstrand

Re: Open Source Is Data Science's Missing Ingredient - by Matt Asay

Glenn Engstrand — Sat, 10 Jan 2015 15:04:50 -0000

An important message, especially for the fortune 500. Open source is so successful in big data that you could not have a credible claim about using big data without open source. You have already mentioned the most common open source projects. What I learned this year at http://it.toolbox.com/blogs... is that there are some additional projects like Cascading, Storm, and Druid that are also worth some attention.

The most significant role in big data where I still see proprietary software being used is discussed at https://about.zoosk.com/zh/... where commercial OLAP cubes are used as a kind of front end GUI to Hive. Other places include connectors to proprietary databases for ETL purposes and cluster management dashboard tools.

Re: Big Data Source Code: Getting Better All The Time - by Matt Asay

Glenn Engstrand — Sat, 13 Dec 2014 21:47:23 -0000

It's hard to measure and compare proprietary code quality to open source code since the source code to most proprietary software cannot be analyzed for errors. I think that most companies today have more trust in open source because of the communities surrounding the more popular projects and because their already trusted engineers can analyze the open source code.

Here are a couple more big data open source projects that you missed. Cascalog makes it much easier for developers to write hadoop map reduce jobs. Apache Spark is a similar technology known for its high performance and has gotten quite a bit of industry attention. See
http://glennengstrand.info/...
for how and why these two technologies accelerate big data engineering. Both Cascalog and Spark are now supported by the major Hadoop vendors Hortonworks and Cloudera.

Re: Hadoop 2.0 & YARN: Get Ready For This Summer's Big Data Breakthrough - by Brian Proffitt

Glenn Engstrand — Mon, 27 May 2013 13:44:52 -0000

YARN, or MRv2, has been out for over a year now. From http://www.dynamicalsoftwar... you can see that it is easy to build Java apps that run inside the Hadoop cluster. As with many open source big data technologies, the sweet spot for ISVs is enterprise integration tools such as ETL 4 OLAP.

Re: Hadoop Adoption Accelerates, But Not For Data Analytics - by Matt Asay

Glenn Engstrand — Sun, 12 May 2013 14:58:11 -0000

I believe that labeling the use of open source software in managing big data as a "poor man's ETL" is a bit disingenuous. Also, I am not so sure about the term "unsupervised landfill" which feels to me to be overly provocative.

I have found ( see http://www.dynamicalsoftwar... ) that companies use Hadoop to crunch their big data down into something that can be loaded into OLAP cubes. Perhaps most companies aren't really aspiring to something as ambitious as a recommendation engine but that is a far cry from "unsupervised landfill."

Re: Open Source Search with Lucene & Solr

Glenn Engstrand — Tue, 23 Apr 2013 02:21:00 -0000

No wonder lucene is so prevalent in so many open source projects. It is high performance yet easy to use. You can find another tutorial embedding the next major release lucene 4 at http://www.dynamicalsoftwar... with emphasis on multi-threading concerns.

Re: Programming and Data Interpretation Are the New Literacy

Glenn Engstrand — Sun, 09 Dec 2012 14:29:40 -0000

Programming and the ability to analyze has always been pretty important in my career. Also, consider http://www.dynamicalsoftwar... for 9 more skills that are a part of basic knowledge worker literacy, especially if some members of your team are remote.

Re: Explorations in Unix

Glenn Engstrand — Mon, 03 Dec 2012 11:17:42 -0000

You might also like http://www.dynamicalsoftwar... which goes into many unix commands (such as grep, cut, join, paste, awk) to reshape application log file data into something that can be analyzed and graphed.

Re: Big Data: What Do You Think It Is?

Glenn Engstrand — Sat, 28 Jul 2012 15:35:39 -0000

I work at a start up who owns the world's largest romantic social network. Like many start ups, big data means the classic three Vs of volume, velocity, and variety. Unlike other start ups, big data (especially social big data) does not mean hadoop. We used http://www.dynamicalsoftwar... to implement our social big data app.

Re: How to Prevent Blogger Burnout: 3 Good Tips

Glenn Engstrand — Wed, 25 Apr 2012 02:46:07 -0000

Good advice, especially the blogging tools. It is still incubating but there is a personal assistant called opportunity eye that helps bloggers discover new topics to blog about that their users have shown interest in. Knowing what will be a hit is another way to prevent blogger burn out.

Re: 4 Blogger Tools for Breaking Your Writing Block

Glenn Engstrand — Sun, 22 Apr 2012 18:37:25 -0000

Keyword based suggestions are fine for a start but, to get better quality topics to blog about, you should get hints based on what your real readers are interested in. That is why you might be interested in the opportunity eye which is a project where your blog readers get to discover related opportunities in a widget surfaced on your blog site. Users are allowed to follow, discuss, or visit opportunities related to the blog topic. Engagement metrics are tracked and published for the blogger to explore.

Re: Hippo Integration with PugPig for iPad Magazine Publishing | CMS Report

Glenn Engstrand — Wed, 07 Mar 2012 01:32:26 -0000

I can corroborate that Hippo CMS is easy to integrate with. During my work on http://www.dynamicalsoftwar... I was easily able to integrate Hippo CMS with Real-Time Communications technology. Congratulations to those folks for taking the time to do JCR-170 the right way.

Re: Alfresco 4.0 to Upgrade CMS for the Cloud

Glenn Engstrand — Wed, 07 Mar 2012 01:24:31 -0000

I was working with Alfresco 3.4 on EC2 during my time with http://www.dynamicalsoftwar... and I can't say that I ran into any cloud related problems.

Re: comSysto and Hippo Team Up To Focus On The Customer

Glenn Engstrand — Sun, 04 Mar 2012 14:41:39 -0000

Hippo CMS has great support for standards compliant inter-operability which puts it in a great position for attracting Enterprise customers. For example, http://www.dynamicalsoftwar... is a beta where group meetings get summarized and published on a Hippo CMS automatically. That was possible only because of Hippo's commitment to JCR.

Re: Alfresco Makes its CMS More Social

Glenn Engstrand — Sun, 04 Mar 2012 14:37:11 -0000

I like Alfresco. It's standards based approach to integration has always allowed third parties such as http://www.dynamicalsoftwar... to enhance Alfresco in ways social and otherwise.

Re: One language everywhere

Glenn Engstrand — Thu, 20 Oct 2011 15:49:56 -0000

I believe this approach to be fundamentally flawed but, if you must, then take a look at clojure.

On a tangent, the difference between developers and engineers is in attitude. A developer believes in the swiss army knife approach to toolsets and wants to show the world how clever he is that he can make his favorite coding language do things way beyond its original scope. An engineer believes in the right tool for the job. Engineers don't identify with any single programming language. Instead, they pride themselves on knowing how to code in lots of different environments and API stacks.

Re: Unlocking Big Data with R

Glenn Engstrand — Fri, 09 Sep 2011 01:20:46 -0000

R is great for engineers and developers who need to analyze big data. Accountants will still want to use OLAP and Excel. If you are more into the data mining part, then take a look at weka. I go into more detail on this at http://www.dynamicalsoftwar... including a short introduction to loading data and generating graphs using R.

Re: Don’t Let Your Babies Grow Up to be Knowledge Workers

Glenn Engstrand — Fri, 25 Mar 2011 01:07:41 -0000

I've been on the other side of this equation as a technical lead in several small organisations who decided to offshore some of their software development responsibilities. Inevitably, the quality of what gets delivered is over promised and under delivered. I attribute the lack to the wide gap in time zone and culture and in the high turn-over that is characteristic of Indian developer shops. Here's the dirty little secret to software development. Engineering quality is the difference between success and failure.

Re: Unlocking the Mayor Badge of Meaninglessness

Glenn Engstrand — Sun, 19 Dec 2010 14:43:33 -0000

Very relevant article. It reminds me of stack overflow where programmers share advice. There is an elaborate reputation system at play over at that franchise. At first, I found their badges to be a motivation but now I find the whole thing to be tainted. I now question the validity of the advice as posters play to the perceived predominant paradigm of the group in order to up their stats.

Re: You’ve Got FMail

Glenn Engstrand — Sat, 13 Nov 2010 17:22:21 -0000

Well, I'm not surprised. Google and Facebook already compete on Instant Messaging (did you know that you can chat with your facebook friends via XMPP?) and they compete on activity streams (Google Buzz vs Facebook Open Stream) so it only stands to reason that email would be next. Maybe Facebook will cook up a collaborative office suite too.

Re: A Beginner’s Guide to Integrated Development Environments

Glenn Engstrand — Wed, 20 Oct 2010 00:42:50 -0000

I cast my vote for Eclipse. If you want to learn more tips and tricks about Eclipse plugins and plugin management, then check out http://www.dynamicalsoftwar...

Re: Ubuntu 10.04 Lucid Lynx - A Fresh Look At What Is On Offer

Glenn Engstrand — Sat, 16 Oct 2010 19:53:22 -0000

Anyone here know when Lucid Lynx users will be able to synchronize their evolution contacts to Ubuntu One Cloud? https://wiki.ubuntu.com/Ubu... indicates that "backporting of the CouchDB and Erlang packages" must happen first.

Re: Eclipse Plugins

Glenn Engstrand — Sun, 03 Oct 2010 14:43:45 -0000

Thanks for sharing your list. I would definitely vote up your top three and will check out the Amazon plugin. Here are some more favorite Eclipse plug ins (see http://www.dynamicalsoftwar... for you to check out.

Re: XMPP servers head-to-head: ejabberd vs Openfire (first impressions)

Glenn Engstrand — Sun, 03 Oct 2010 14:40:11 -0000

If scalability is key, then check out Tigase which is a highly scalable XMPP server written in Java. I have compared Tigase and Openfire (see http://www.dynamicalsoftwar... including notes on how to write extensions. Tigase is just as easy to extend as Openfire and offers more ways in which to extend.

Re: XMPP becoming common standard for real-time chat applications

Glenn Engstrand — Sun, 12 Sep 2010 12:35:58 -0000

Thanks for this interesting introduction. Can you be more specific about the XMPP interface for Facebook?

You mentioned jabber and open source. Both openfire and tigase are two modern open source XMPP server projects that you might be interested int. http://www.dynamicalsoftwar... is an article that compares and contrasts those two technologies.

Re: Yammer 2.0 To Launch As A Powerful, Full-Fledged Social Network For The Enterprise

Glenn Engstrand — Sun, 05 Sep 2010 18:53:03 -0000

We use yammer at work. It's great at broadcasting changes (kicking a server, upgrading a package, announcing a new service) to interested parties in the company. An activity stream is a good way to make announcements that are searchable for those who are interested without the heavy digital footprint of email. Yammer makes a lot of sense; however, the web interface does not. It's too much of a distraction. That's why I use Gabble which brings out the best of micro-blogging while filtering out the excessive noise of a full web portal.

These specialized clients are the way to go for real time communications. That's why I use Gwibber for micro-blogs and Empathy for IM. XMPP is for real time what HTTP was for publishing.

RTC is not just for human consumption. Check out http://it.toolbox.com/blogs... for an example of how these advances help the enterprise in other communication intensive activities.