No Statisticians were Harmed in Making this Graph - Extrapolating GlassFish Trends

ALT DESCR

I felt like a break so I took the latest Google Trend on searches on GlassFish, WebLogic, WebSphere, JBoss (live, archived) and tried to apply a smooth curve on it. No statisticians were harmed in eye-balling these graphs... but the cross-overs are WebLogic - 2010 and WebSphere and JBoss - 2012.

If a Statistics-savy guys wants to give it a try please let me know what you get :-)

We are in the middle of a substantial upswing on the GF; the world economy "troubles" seem to be accelerating our adoption significantly so I suspect the data will become much clearer in a few months.

Comments:

Eduardo, nice chart. One minor change I think would be fair is to qualify each term with "Java" as GlassFish, unlike the others, means something outside middleware. I don't think it changes much. For example add Geronimo to the chart with and without the Java qualifier - see what I mean.

So here's my version - http://tinyurl.com/3uj7jh

It will be interesting to look back in a few years time. I expect the increasing OSS momentum will change the chart significantly.

Posted by Rich Sharples on December 18, 2008 at 09:16 AM PST #

Here's a similar one for OpenSSO and some of its closed source competitors: http://is.gd/cqNJ [Google Trends]

Just give us a few more months... :-)

Posted by Pat Patterson on December 18, 2008 at 09:30 AM PST #

To Rich --
re: I expect the increasing OSS momentum will change the chart significantly -- yep, I expect that too. I do think the next six months will be quite interesting.

re: adding "java" -- Yes, "GlassFish" is less unique than "JBoss", but not by much. I looked at the first 100 hits by google and only 7 were not "our" GlassFish (I made sure I was logged out of google to avoid personalized query artifacts). That is very different than "Geronimo" or "Tomcat", which have high hit rates outside of the server containers.

Adding "Java" to all the queries changes the results but it is not necessarily the case that it modifies them equally. Recall that trends tracks searches, not hits. Given that "GlassFish" is pretty unique, I can't think why anybody would search for "GlassFish Java", and it certainly changes the growth rate very substantially, see http://tinyurl.com/3jakgr.

As you say, it may not change much, but the sample size will be much smaller, and thus, possibly less representative. Since I've been tracking "GlassFish" on its own for quite some time now, I'll continue to do that.

In any case, all these metrics have limitations, so I only use them as indicators. Fortunately they are all pointing the same way :-)

- eduard/o

Posted by Eduardo Pelegri-Llopart on December 18, 2008 at 01:44 PM PST #

I think you did a fair job of smoothing the lines and extrapolating the data. But there is something missing. Most successful new products will hit a tipping point where adoption rates increase. So my prediction is the line for GlassFish will get steeper at some point in the future although I have no idea when. Look for a hockey stick.

Posted by Ed.T on December 18, 2008 at 09:41 PM PST #

EdT - We might be very close to the hockey stick. There was a strong pickup
with GFv3 and NB 6.5 but then we got into the usual ThanksGiving drop (see
http://tinyurl.com/3t9w7f ). Except that in previous years the drop continued
through Christmas and we didn't recover the volume of searches until mid-year,
and this year we have already recovered.

We are also showing in places like the Devoxx whiteboards (see
http://tinyurl.com/4lhxba ). And internally our sales guys are hitting us with
continuous requests, but I can't talk about that in any detail.

There are so many variables right now that's hard to do a good prediction. I
personally believe we will see a strong acceleration in February.

- eduard/o

Posted by Eduardo Pelegri-Llopart on December 19, 2008 at 12:17 AM PST #

Post a Comment:
Comments are closed for this entry.