Three Years to overhaul a website - a failure?
November 3, 2009 · 12 Comments
TechCrunch ran a story yesterday about MFG.com's site overhaul. The new site uses Java and took three years for their team to complete, during which time the team felt "hand-cuffed" and the "great ideas" the company had couldn't be executed.
To me, that whole story sounds like an expensive failure.
Of course, the fact that it was a ColdFusion to Java transition has drawn a lot of comments from the CF community but most of those comments focus on the CF side of things and the suspicion that MFG.com didn't do its research (I personally suspect they got hooked in by an "enterprise" consulting group who started them down the Java tracks).
Whilst we can all sing the praises of CFML and criticize MFG.com for many things, I think we all need to look at the business aspects of a three year project to relaunch a web site in this fast-paced world. Yes, I'm sure it was a massive project and involved a lot of integration work but even so, macromedia.com was completely rebuilt from the ground up in less than that - and Macromedia's complete ERP replacement project (with 38 integration points!) was also completed in substantially less than three years...
Tags: coldfusion · j2ee
12 responses so far ↓
1 Sami Hoda // Nov 3, 2009 at 6:30 PM
2 Spills // Nov 4, 2009 at 8:26 AM
3 David Tannersyn // Nov 5, 2009 at 5:10 AM
I spotted your comment on the Tech Crunch, MFG.com article (among many other CF advocates), found your blog, and had a look around. It is clear that you have a much more solid programming and architecture background than I have in diverse development languages. I feel compelled to ask you a question after reading the some-what over the top comments from the CF community. I hope you do publish this post as it would be great to get more feedback from the type of people that would read your blog.
I have worked in a number of software houses, and also have some experience in asp.Net, Java, and a good number of years in CF. In the past there was no doubt that CF was the avenue for a quick time to market, scalable and fast site .
My question/thinking:
In Nov 2009 why recommend Coldfusion in good conscience for a new project or a career when Job postings for ColdFusion developers are remarkably low compared to Java and .net, it will unlikely be found on a university syllabus (is it taught anywhere?) while Java and .net are ubiquitous, and has even a lower take up outside of the USA (about zero take up in non-English speaking countries) where Java is thriving. This all seems to be the symptoms of a dying technology. The job postings are very, very key for those starting new projects or careers. Also, the frameworks and capabilities of CF are getting so similar to Java why not use Java, and ensure you have resources to work on your project in the years to come?
My assumptions:
- I am focusing on very dynamic, constantly changing websites that are there for the long haul, and development teams greater than 2 resources that will have rotating over time - the site will need a framework, coding standards, releases processes, and scheduled refactoring i.e. real, maintainable websites (the goal of any website regardless of its start).
- The level of experience of the CF developers and, say, the Java developers are at an intermediate to senior level. They know their stuff and have a good solid background in the type of website above i.e. real, experienced programmers who are concerned about processes and standards that will see the site through a number of years (the goal of any developer regardless of their language).
Thank you,
David Tannersyn
4 Peter Bell // Nov 6, 2009 at 9:51 AM
There are still substantial productivity benefits in ColdFusion over Java or .net. If you were really concerned about ColdFusion, you would do better to look at Groovy or Ruby.
Ruby is still an outsiders choice in large corporations and there is still an overhead in keeping up with the quickly evolving deployment infrastructure.
Groovy is excellent. The productivity of a dynamically typed scripting language but deployment through standard WAR files. It's also really easy to integrate Java and Groovy if you need better performance in certain aspects of the application.
With ColdFusion (especially Railo - an open source CF engine which to me competes squarely with Groovy) you also get the productivity benefits and it's a much more mature infrastructure. Today there are probably more CF devs than groovy devs and you don't have to work around the bugs that still slow down Grails development.
You could also look at Python, but I think despite Googles support that will lose ground, or newer languages like Scala or Clojure, but they're a little esoteric for regular business web apps unless you need to handle concurrency issues or love lisp style meta-programming.
5 Sean Corfield // Nov 6, 2009 at 11:20 AM
6 James Mohler // Nov 7, 2009 at 10:52 AM
1.) Its not Legacy technology. Non developers need to know that it isn't a Legacy
2.) Let's talk about the aspects of the CF that are better. C# have struggle with things that CF can do without any effort.
3.) Some one many years ago showed me CF. I need to show it for others. That way they know there is more out there than ASP.Net and Java
~~~~~~~~~~~~~~~~~~~
Slightly off topic. The .Net developers I work with are blown away by <cfdump>, <cfwddx> and query of queries.
7 Brian // Nov 12, 2009 at 11:13 AM
8 David // Nov 23, 2009 at 5:46 PM
There is no conceivable way you can tell me that the business assumptions, and technology options available on day 1 were the same as on day 1095. IT has been on a massive growth phase over that time.
The reason why the CF community is passionate about CF is because the deep tool set and rapid development options allow us to reduce the lead times between business problem and business solution.
There are NO technology problems, there are ONLY business problems. ColdFusion allows us to concentrate on the former, and not worry about the latter.
9 walt // Dec 13, 2009 at 8:16 AM
just checking to see if you'd had time to post a response to David's post. I'm very interested to read it.
Thanks!
10 Sean Corfield // Dec 13, 2009 at 8:44 AM
11 walt // Dec 13, 2009 at 9:24 AM
:D
I'm interested to read it, as I'm a 4/5 year CF developer, and I seem to be a rare commodity here in Dallas. (I may be completely off base, just going by what I've seen in my own experience)
Most other CF developers I've met have 6 years + in CF coding.
Anyway, no rush, I'll keep an eye out for it when you do have time for it.
12 Sean Corfield // Dec 13, 2009 at 10:29 AM
http://corfield.org/entry/Why_CFML_Why_not_Java
As for being a rare commodity in Dallas - I don't think you are... My experience is that area has a pretty active community and a lot of job openings (certainly I've seen a LOT of jobs posted in that area lately!).
Leave a Comment