Humanware?

Is Fundación Chile sending SPAM?

I got an email that supposedly comes from Fundación Chile, though it’s actually sent by imacom, with an “invitation” to a seminar on HumanWare.

Besides the fact that I don’t like SPAM, and that I hope this time the people at imacom really do remove me from their lists, I want to take the opportunity to save you CLP 320,000 (550 dollars), which is what the seminar costs.

Sure, because they will most likely spend a good couple of hours proving to you that the IT industry is in crisis. The introduction to the seminar’s presentation cites the much-questioned Standish Report. If we start from a foundation that weak, I doubt the rest of the seminar is of any real use.

The criticism of the Standish Report

Robert L. Glass, in the August 2005 issue of the famous Communications of the ACM, under the title “The Standish report: does it really describe a software crisis?”, asks whether this report, cited so often, really describes a software crisis. Glass asks us to reconsider how relevant the report is.

The Standish Group’s report, first published in 1994 under the title The CHAOS Chronicles, sets out the most quoted statistics on the subject, namely: “49% of projects finish late and over budget. Or 23% of projects are canceled before completion. To the point that, according to Glass, the idea of a permanent software crisis has been perpetuated.

Glass is a respected expert in the field. He has been working in the industry for some 45 years, he is an academic, he was named an ACM fellow in 1998, and he has written several books on the subject.

Glass asks a very simple question: how can this crisis exist in “an era that simply would not be possible if we weren’t so successful at building software that makes computers do all the wonders they do?” Clearly something doesn’t add up for Glass.

In the ACM article Glass explains how it has been impossible to get the Standish Group’s data; what we have is data from 1994. Reports that criticize the Standish Report are rarely cited. An important 2004 report by the Simula Research Laboratory in Norway questions the fact that in 1994 the Standish report said projects overran their schedule and budget by 184%, and that these figures have been going down in the more recent reports.

Have companies gotten better at estimating their own performance? In its report, the Simula group wants to establish that continuing to cite the Standish Report is not only questionable in terms of validity, but that the continued use of its results holds back progress in the field.

Why is the Standish Report harmful?

The Simula group gives real-world examples that help make the case:

Example 1: A project was 146% over budget, that is, it cost 2.5 times the original estimate. A report on the project’s performance stated that this cost overrun wasn’t so bad, because it was much better than the industry average (189%).

Example 2: A consulting firm presented itself as an elite software development shop because it compared its own numbers against the 189% overrun given by the Standish Report.

But the worst part is that the definition of what a 189% cost overrun actually means was something the Simula researchers had a very hard time finding. The Standish Group’s handling of the figures is pretty slippery, to say the least. The Simula group’s report is out there; you can see for yourselves.

It looks as if the entire Standish Report is sliced to show negative results. Many figures exclude successful projects. Nor is there a category for projects that come in under the estimated budget, and not only in the Standish report.

It’s as if we IT people were used to flogging ourselves.

Humanware?

I won’t go on; Glass has written plenty criticizing the Standish Report’s numbers. I recommend reading his article published in Computer magazine in March 2005.

What bothers me is that these figures and dubious studies, from one specific field, get used to introduce neologisms, to confuse, and to wrap in new paper the same old candy, which has no flavor left from being handled so much.

Humanware: the seminar in question delivers “the tools needed to prevent, diagnose early, and/or manage the conflicts inherent in a Technological Change project”.

What the hell is Humanware? What does the Standish study have to do with managing human resources, in any project?

Could it be that what’s missing is ThinkWare and what we have too much of is ConsultantWare?