Manual processes can be better

by Lisa Nicholls Sun, April 03 2011 18:07

Thomas Dolby didn't use any auto-tune on Oceanea.  As you'll find out if you follow the link in the last sentence, however, he does go in and electronically re-tune individual syllables wherever and whenever he feels like it.

Well, of course he does it "electronically".  He's Thomas Dolby, for the love of mike.  How else would he do it?  (Come to think of it how else would anybody re-tune a syllable? I haven't a clue.) 

In the music world, a lot of people think of auto-tune (pitch-correction software) as a way of cheating.  If your voice is not exactly true on a note, you can let the engineers fix it... but it's not the same as having a great voice.

While software developers often have a lot in common with musicians, in this case they often take the opposite view, especially in the enterprise-class software world:

Automating every phase of the process (from class generation, to code review, to build-and-deploy, and beyond) is the goal.  If you haven't automated something, you should be figuring out how to fix that hole in your process.

Automation = time saved = overall productivity.  Sure, it takes time and effort to figure out, build, and get used to an automated process... but if you really care, you'll put in that time and effort, because the result will be better.

I've made that same argument myself, numerous times, in numerous scenarios.  But I think we need to take a note from Thomas Dolby here. 

He's not using a manual process because he doesn't care.  He's using a manual process because he cares that much. He can do better, by hand, than he can do using auto-tune.

And, sometimes, we can do better too.   Sometimes we use a manual method because we think we can do better than the automated tools.

I suppose I should add "... and when that happens, we don't care how long it takes to do better."  Well, okay.

Time saved != productivity, except in the grossest sense.  Productivity, by the way, != excellence.  Productivity is only one component of excellence.

I can hear the obvious objection...

"What you're talking about won't scale, Lisa". 

You know what?  Sometimes, scaling has nothing to do with excellence either.  When that happens, scaling really is something I don't care about.

Tags:

General TechSpeak | Life

Comments (5) -

5/25/2011 5:43:50 AM #

Portrait Artist

Interesting view with which I absolutely agree. When resorting to manually proceed means that we want to put more of myself. To show some artistic design

Portrait Artist United Kingdom

7/25/2011 9:32:36 AM #

LSN

Tony makes a good point.  

IMHO it's not as cut and dried as "manual process is always better"... and not even as cut and dried as "manual process is better for dev phase, automation is better for stable software". But Tony's distinction is a useful triaging tool.

LSN United States

6/19/2012 11:27:49 PM #

Leonard

Automation doesn't mean 'you cannot see how it works step by step', you can always learn something if you really want to.
So, in the software area, I think this is not about automation making people know nothing but whether have necessary to truly and deeply understand everything.

Once more thing, if we can do better than the automated tools, sometimes it means the automated tools can be improved, since automated tools is still be invented by the software engineer. I guess you will say 'Inventing a automated tools is a manual process', <s>

Leonard People's Republic of China

8/3/2012 9:52:20 PM #

&gt;L&lt;

The problem, Leonard, is that people use automation as a way not to think <sigh>.  It's not that automation is bad in itself, it is that it allows people to believe meaningful work is going on simply because there are a lot of lines of code and a lot of activity.  

There may be many generated unit tests, for example, without anybody thinking about what really needs testing.  People develop a false sense of confidence because there are so many tests, they must have done a great job testing when in fact they haven't tested anything important at all.

"Inventing an automated tool" may in fact be a manual process, or it is actually possible to generate tools (I've done this!).  But "inventing" the tool and using the tool intelligently are two different things.  The process of inventing a tool is fun and interesting to me but the use of the tool to get real work done is the most important thing.  Tools are fun and important, but inventing them can get in the way of actually performing real work, it can be a distraction.  And using them after they are invented can be more of a distraction, and more of a detraction from true quality.

>L< United States

11/1/2012 4:23:16 PM #

Tyler Martin

Scaling is all most people seem to care about these days. I guess it is a sign of our ever accelerating culture. We want to get more done in less time and max out our results and profit (profit in every sense of the word).

But man oh man sometimes we just need to return to manual hands on, FOCUSED & ATTENTIVE action and work.

The best thing is that these things can feed into each other. Doing something manually and seeing the value that comes from it can illuminate those processes which can and should, and could be automated. Likewise automation can teach you which processes need to remain manual and need more care and attention.

It's the balance that counts, but you need to be manually attentive to see that balance.

my brain hurts.

Tyler Martin Canada