The 80/20 Rule and Nested Sub-projects

ML Jan 26, 2006

OK, I’m starting on these 2 projects, but I’ve got the documentation bug. These are two very mainstream projects, that would be of great use to the world at large. And I’m going to build them up from scratch, using nothing more than what’s already installed on most people’s desktop PCs. So, I want to document it with my baby-step technique, where I show the entire program as it develops on every page of the tutorial. It’s quite inefficient from a file standpoint. But if the CMS system makes it manageable, there’s really no harm. It is the Web after all, and you don’t have to kill trees for more pages. And if it makes the student’s experience better, it’s worth it.

But that leads to a nested sub-project, and the issues of whether or not to do it. I am a big believer in the 80/20 rule that states you should plan to get 80% of the benefit of any endeavor from the first 20% of the work (there are other interpretations for tax purposes, etc.). So, a series of half-implemented projects has a net gain and still moves you forward. Nested sub-projects which plague many professions are the enemy of productivity and the 80/20 rule. Suddenly, you’re embarking on project after project before you even start on the first thing. I even wrote an 80/20 rule poem.

You saw an example of me avoiding a nested sub-project pitfall when I decided to just go ahead with Blogger. I could have tried installing WordPad, had another server and database to deal with, a new system to learn, etc. I could have even written my own (which I have partially working). But by choosing Blogger, I could just forge ahead. And here I am the next day with many blog posts under my belt and standing at the edge of another potential pitfall, looking over the edge. Let me explain.

I can already do baby-step tutorials using my CMS. The problem is that to make them really cool, you have to highlight what lines changed in the code from page to page. Manually inserting the markup is tedious, and defeats the purpose. It can and should be automated, and I already used a Windows derivative of the Unix diff program to get started. Its half-way implemented. I basically make a post using my home-grown blogging system, and it looks a the previous posts, finds the differences, and automatically inserts highlighting and strike-out code to show on the current post what changed from the previous post. It’s way-cool, and the foundation for something new and innovative on the Internet in its own right. I could easily imagine this site becoming as popular for the novel baby-step tutorials as it does for the MyLongTail app. Problem is, it’s only half-implemented, and I don’t know whether I should try to bang it out all the way before today’s main projects.

Let’s evaluate. Ask the key questions…

  1. Can you possibly imagine even more nested sub-projects? Or is it a clear one-off? Will you get caught in a recursive trap and a nightmare of maintenance overhead?
  2. Is it foundational, meaning that it will improve all the other work you do and start resulting in compounding returns? So, over time, do you really save time?
  3. Is there a better already existing way to do this?
  4. What are the corollary benefits, and do they outweigh the lost time?
  5. Are there urgent aspects of the projects you’re putting aside? What is the damage of delay?
  6. Is it really necessary for your primary objectives?

This should be a clear one-off project, because it is basically a rebound-action on a database insert. When the insert occurs, just do this quick processing. The auto-markup occurs.

It is foundational, but only on the documentation side. If you consider documentation foundational. Its way different than other documentation systems in that it captures process, for better or for worse. It is much like the Disney work-in-progress pencil-sketch test animations that have more character than the finished product. It also gives you more ability to learn about the animation process than the finished product. It is rare to the point of non-existence in the programming world, because it takes too much time to document in this way, and reveals the many imperfections of the creative process (because it documents mistakes and all). The closes thing to this is Wiki revision tracking, and code version management software. Anyway, all this is to say, yes, I believe the work is foundational, because it is key to providing a rich documentation and tutorial experience on this site.

This gets to the fact that I already made the decision to use my home-grown CMS system. With that decision made, I need to choose something that integrates well. And I actually already am choosing “the better way” to do this in tying in the Unix-derivative diff program. I could have attempted to actually program this from scratch, but this program gives me everything I need to parse a file and insert the code. I can focus on parsing and marking up instead of detecting differences.

There are massive corollary benefits. It allows me clearer differentiation of what I document in Blogger, and what goes in the CMS (stream-of-consciousness goes in Blogger, and baby-step code tutorials go in CMS). The tutorials increase coolness factor and buzzworthiness the chances of getting this site written about, and eventually SlashDotted. That is not only a corollary benefit, but is a main objective. The project also clarifies my own thinking, making me code more carefully knowing that the non-proprietary parts are going to be public and under the scrutiny by other programmers.

Yes, there is a very urgent aspect of the projects I’m putting aside. I want to document the very first search hits to ever occur on MyLongTail, and the very first GoogleBot visit. I may miss them. The site is already out there, and I’ve been blogging (but without pinging). Will the site be that much less interesting if I miss these key events? If I can really isolate the projects (all 3) down to a single day, will I be really jeopardizing it that much more? I don’t think so. So, all three projects should be done in one day. But one of those projects is really less important than the others. More on that soon.

While not necessary to my primary objectives, it certainly does help the “erupt in buzz” objective. More tutorials mean more pages of real value to a broader audience, and more search optimized pages, and more pages using a unique and valuable tutorial technique that perhaps the buzz-brokers will recognize. It’s also worth noting that I already have the blogging bug, which is somewhat cutting into just getting the work done (its 11:22AM already). And having this system running will let me feed that blogging appetite while simultaneously actually doing the coding. So, while not necessary for my primary objectives, it highly reinforces them, and I will move ahead with the baby-step tutorial marker upper.

Leave a Reply

Your email address will not be published. Required fields are marked *