Search and the Open Source Card
February 22, 2010
A happy quack to the reader who sent me a link to Michael Tiemann’s “How Open Source Software Can Save the ICT Industry One Trillion Dollars per Year.” You can find the seven page document at http://regmedia.co.uk/2010/02/18/tiemann_cost_of_development_paper.pdf. When the paper was written in the fall of 2009, Michael Tiemann was the President Open Source Initiative and Vice President Open Source Affairs, at Red Hat. This firm is one of the highest profile commercial enterprises to have built a business on open source software. You can get the rosy financial news by searching Google Finance for RHT.
When I read the paper, I found myself in general agreement. But Red Hat is in the operating system and middleware business. For companies eager to chop down the license fees charged by commercial software companies, Red Hat’s approach is a must-have play.
My interest is search and content processing, and I think that many organizations are struggling to define search. If the news flowing from companies like Lemur Consulting and Lucid Imagination is accurate, some commercial search vendors no longer get a chance to compete. The outfits happy with Red Hat, JBoss, and other open source software are likely to hop on the Lucene / Solr bandwagon.
You can get a very upbeat picture of the benefits of open source software in Mr. Tiemann’s white paper. So if you want to make a case to go open source, you will want to download the document and tuck it in your “Sources” file.
There are some interesting “factoids” in the paper; for instance:
- A reminder that most commercial software installations end up as train wrecks.
- Costs and unnecessary expenses continue to escalate for organizations relying on commercial software
- Proprietary software inhibits innovation.
But what about search?
Let me identify what I think is an interesting trend regarding open source and commercial vendors of search and content processing systems.
First, I have noted that one company has cut a deal with a commercial enterprise to make “connectors” available to the open source licensees. Connectors are the code widgets that allow one type of content such as Lotus Notes email to be indexed by a third-party system such as Lucene. This merging of commercial and open source suggests to me that for certain types of software, the open source community does not provide what many organizations need. After all, what good is a search system if it cannot index information in a widely used email system like Lotus Notes? I am not suggesting that the rosy picture painted my Mr. Tiemann is incorrect, but I think this is an interesting open source gap. Perhaps it will be filled by Red Hat?
Second, a number of high profile companies are offering open source operating systems. One notable example is a large search vendor’s operating system for mobile devices. If I were a struggling mobile company, I would certainly look closely at an open source, no-fee operating system. One would think that such a mobile operating system would sweep through the telecommunications industry like wildfire. What I learned last week was that Motorola was giving the for fee Windows 7 Mobile a very close look. Why? If the open source mobile operating system has a fraction of the payoffs referenced in Mr. Tiemann’s essay, why hook up with a very proprietary outfit like Microsoft? What does Motorola know that I don’t know?
Third, a number of vendors are talking about such Frankencode approaches as “support for open systems”, “full embrace of standards,” and “our APIs are open”. What do these phrases mean? On the surface, these vendors of proprietary systems seem to be leading me down the open source path. However, are these vendors using language in a way to lure the red fox to the steel trap?
Fourth, a very large outfit has figured out how to run Linux on its mainframes. What’s the purpose of this technical cartwheel? If I “buy” a mainframe, won’t the margins be sustained by boosting the price of those funny little connectors that mainframes use to hold drives in the DASD or the truly weird cables needed to hook certified gizmo A to certified gizmo B?
My hunch is that open source is a significant trend in software. Some of the success of open source is driven by those who want to create software to hold down costs and operate in a manner that to some degree reduces the brutal costs associated with certain commercial software products.
I think there is a big marketing and PR play underway as well. The use of the phrases “open source” and “support standards” sounds pretty good. Get the software into the company. When the organization’s boss figures out that the existing tech staff cannot make the open source software work as everyone believed it would, then the consulting engineers are ready to pounce.
My view is that one needs to bring the same discipline to defining requirements, testing software, and performing financial analyses regardless of the software type. This means that commercial and open source adherents will have to prove that their products and services can stand and deliver.
Without that discipline, “open source” is little more than a buzzword like “social media.”
Stephen E Arnold, February 22, 2010
No one paid me to write about open source. Because open source is “free” and I was not compensated, I am at a loss to know to whom to report my financial lapse. Maybe the Department of Treasury is the outfit in charge? Treasury knows money or at least how to print it I believe.
Comments
One Response to “Search and the Open Source Card”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
Hm, lots of stuff here, Stephen. Where should I start…. Here are some quick reactions:
* Yes, there are organizations who push open source mainly to lower costs or undercut more closed-sourced competitors. I’d say there are still more individual enthusiasts. This may change over time.
* “Open APIs” and such claims from vendors – I often don’t know what this even means! Fluff.
* “When the organization’s boss figures out that the existing tech staff cannot make the open source software work as everyone believed it would, then the consulting engineers are ready to pounce.” – I’d say that’s the same regardless of which software is chosen. I just spoke to somebody last week, a person from a very large bank we all know, who called Sematext because they couldn’t implement something with FAST. Open-source ecosystem has consultants, closed-source companies have professional services.
* Speaking of fast (and Convera), I think it’s a golden opportunity for many organizations to finally give Lucene/Solr a try: http://blog.sematext.com/2010/02/13/run-fast-to-open-source-search/