The Old Joel on Software Forum: Part 4 (of 5)

Choose Two

I just spent a bit of time reading this.

http://www.xulplanet.com/ndeakin/arts/twelve/

(thanks for the xulplanet link -fool for python- !)

Some of the thoughts expressed (usability, tradeoffs, etc.) remind me of Joel's writings. I'm wondering if the author was inspired by Joel or they had parallel evolution of thought. Either way I found the "twelve days" to be well thought out and concise. As I read I realized that the failure of the last project I was involved in was largely due to the fact that it was attempting to meet almost all of the following

* Efficiency
* Functionality
* Reliability
* Usability
* Customizability
* Extensibility
* Localizability
* Accessibility
* Portability
* Compatibility
* Security
* Marketability

but as the author notes, for the 1.0 you only get to pick two. Remarkably similar to Wexelblat's Scheduling Algorithm:

Choose two:

* Good
* Fast
* Cheap

Is it common for this kind of over-extension (or lack of focus) to ruin a project and what suggestions do you have to focus a team on such a limited area of the software?

lumberjack
Tuesday, March 16, 2004

The link you posted gives a 404 error.

Ken Klose
Tuesday, March 16, 2004

Remove the trailing space.

Karel Thönissen
www.hello.nl

Karel Thönissen
Tuesday, March 16, 2004

Surround URLs with spaces.

but not on the end I guess???

i'll try again

http://www.xulplanet.com/ndeakin/arts/twelve/

lumberjack
Tuesday, March 16, 2004

"Is it common for this kind of over-extension (or lack of focus) to ruin a project"

Lack of focus ruins far more than just software projects. When you don't know what's important, everything seems important. That's why you have to find out where your leverage is. Read "The Goal" by Eli Goldratt. It's all about how to exploit your leverage points and subordinate everything else to them in pursuit of your goal.

contributor
Tuesday, March 16, 2004

An interesting article but I don't agree you are limited to just two for an initial release. Why would that be so? Is there some tradeoff between functionality and marketability? No, of course not. The author is just randomly saying the number two for no reason.

Nearly all first release software we have issued hits 7 of the 12 points.

Some company that only produces 2 of those 12 for a first release is doing EXTREMELY poorly.

Shrinkwrapper
Tuesday, March 16, 2004

Shrinkwrapper,

My guess is that you can only pick two that will differentiate your software from the competition. At the same time you must adequately meet the other 10.

contributor,

Me asking "Does lack of focus ruin projects?" was not really my intention. My typing and thinking were out of sync. I'm really most interested in suggestions ("The Goal") or experiences regarding team direction and tradeoffs. Thanks for the recommendation.

lumberjack
Tuesday, March 16, 2004

I am with shrinkwrapper on this one. Picking two seems very random.

A lot of the choices are not mutually exclusive....

Tapiwa
Wednesday, March 17, 2004

> A lot of the choices are not mutually
> exclusive....

You're right, they're not. But your goal on a project can't be to excel at all of them.

Meaning, you must choose where you will focus your resources. If you've got enough resources to push these 12 points to the max, great for you - and not so great for your competitors :)

However, as MS, IBM, et al have shown us so often, no one has enough resources for this.

-----
"Suravye ninto manshima taishite (Peace favor your sword)" (Shienaran salute)
"Life is a dream from which we all must wake before we can dream again" (Amys, Aiel Wise One)

Paulo Caetano
Wednesday, March 17, 2004

XProgramming has the same type of metaphor, from a business perspective.

The metaphor is that you have four big levers on the software machine. They are Cost, Quality, Scope and Time. However, it's not a matter of choosing two, it is more that all of the levers are interconnected, and as you move one, the others move too.

So lets say the customer wants to reduce time, perhaps to meet a deadline - well, either the cost lever is going to go up (more people or better tools) or the scope lever has to go down. The others work in a similar way.

It sounds like obvious, but we've found it to be excellent in bringing customers down to earth.

Rhys Keepence
Wednesday, March 17, 2004

In my experience of the whole list the top three of "items most likely to turn your project into a mess without any real benefit" are by far:

* Customizability
* Extensibility
* Portability

My top three candidates for success:

* Functionality
* Usability
* Marketability

Just me (Sir to you)
Wednesday, March 17, 2004

Sir,

I think you're basically correct but in some projects customizability is one of the marketing requirements.

On one of our products, we had to through away the first generation product and basically start over because it wasn't customizable enough.

The second generation product was customizable by design, but we did take a bit of detour due to the idea it was going to be portable too.

pdq

pdq
Wednesday, March 17, 2004

Thanks Rhys, I'm pretty sure I'll use "four big levers" line at some point in the future. It sounds like a good way to bring a dose of reality to overly enthusiastic requirements by clients, managers, etc.

So what would the top two for an IDE be?

Possibly

* Usability
* Extensibility

While trying hard not to throw marketability and friends out the window.

lumberjack
Wednesday, March 17, 2004

"* Extensibility"

I'd argue that Extensibility is of almost no value in v1.0, in most cases. Unless the product was some sort of "wrapper" who's only function is to BE extensible (like some sort of DLL builder)

Extensibility only has value if v1.0 has some some value to extend.

I.e., at the extreme, if I produced an amazingly fast program that didn't actually do anything, but was amazingly extensible, why would anyone extend it?

(Wait, isn't that a description of the sw program Chandler?)

Mr. Analogy (formerly The real Entrepreneur)
Wednesday, March 17, 2004