Archive
Users of Open source software are either the product or are irrelevant
Users of software often believe that the people who produced it are interested in the views and opinions of their users.
In a commercial environment there are producers who are interested in what users have to say, because of the financial incentive, and some Open source software exists within this framework, e.g., companies selling support services for the code they have Open sourced.
Without a commercial framework why would any producer of Open source be interested in the views and opinions of users? Answers include:
- users are the product: I have met developers who are strongly motivated by the pleasure they get from being at the center of a community, i.e., the community of people using the software that they are actively involved in creating,
- users are irrelevant: other developers are more strongly motivated by the pleasure of building things and in particular building using what are considered to be the right tools and techniques, i.e., the good-enough approach is not considered to be good-enough.
How do these Open source dynamics play out in the evolution of software systems?
It is the business types who are user oriented or rather source of money oriented. They are the people who stop developers from upsetting customers by introducing incompatibilities with previous releases, in fact a lot of the time its the business types pushing developers to concentrate on fixing problems and not wasting time doing more interesting stuff.
Keeping existing customers happy often means that products change slowly and gradually ossify (unless the company involved has a dominant market position that forces users to take what they are given).
A project primarily fueled its developers’ desire for pleasure cannot stand still, it has to change. Users who complain that change has an impact on their immediate need for compatibility get told that today’s pain is necessary for a brighter future tomorrow.
An example of users as the product is gcc. Companies pay people to work on gcc because they want access to the huge amount of software written by ‘the product’ using gcc.
As example of users as irrelevant is llvm. Apple is paying for llvm to happen and what does this have to do with anybody else using it?
The users of Angular.js now know what camp they are in.
As Bob Dylan pointed out: “Just because you like my stuff doesn’t mean I owe you anything.”
Free range software developers: Are they cost effective?
I have just been reading an eye-opening article by Ramin Shokrizade about the techniques that online game designers use to extract money from players. Playing computer games and writing software have a great deal in common, two important characteristics they both share are being immersive and very enjoyable.
Reward Removal
From the article: “The technique involves giving the player some really huge reward, that makes them really happy, and then threatening to take it away if they do not spend.” Hmm, this sounds familiar. Beginner programmers are very resistant to deleting any code they have written, whereas more experienced developers are much less resistant to deleting code but they often put up a fight if an attempt is made to remove a feature they are responsible for creating.
“The longer you allow the player to have the reward before you take it away, the more powerful is the effect.” Wot! Remove this feature? What if somebody somewhere is using it?
“… uses the same technique at the end of each dungeon again in the form of an inventory cap. The player is given a number of “eggs” as rewards, the contents of which have to be held in inventory. If your small inventory space is exceeded, again those eggs are taken from you unless you spend to increase your inventory space.” Why are there no researchers with this kind of penetrating insight investigating how to make software engineering more cost effective? We continue to suffer from the programming is logic by other means world view, promulgated by the failed mathematicians that populate so many computing departments.
Premium Currencies
“To maximize the efficacy of a coercive monetization model, you must use a premium currency, …” [a premium currency is in-game money that is disconnected from real wold money]. The lesson here is that if you want software developers to make decisions relating to real world events you need to provide a direct and transparent connection to the real world. Hide the connection under layers of abstraction or vague metrics and developers can be easily fooled into making poor decisions.
Skill Games vs. Money Games
“A game of skill … ability to make sound decisions primarily determines … success. A money game … ability to spend money is the primary determinant of … success. Consumers far prefer skill games to money games, …. A key skill in deploying a coercive monetization model is to disguise your money game as a skill game.”
I think most developers consider their job to be one of making skillful decisions rather than one of making money for their employer, rationalizing that these skillful decisions result in their employer making money. Hmm, how much time do developers spend in skillful activity for what appear to outsiders as obscure coding issues; skillful activity is enjoyable while doing what makes most money for one’s employer can result in having to do lots of really dull and boring tasks. I cannot help but think that skill here is playing the role of a premium currency.
The big difference between playing a game and writing software is that in most cases a game has a well defined ending, a path exists to get there and players know when they get there. One of the reasons that managing software developers is like herding cats is that the ‘end’ is often very fuzzy and ill-defined. This does not mean that factory farming techniques are not applicable to software development, just that we have not yet figured out which techniques work.
Recent Comments