Weblog Without Honor or Humanity

Programmers don’t steal enough.

Now that I have your attention: what I really mean is that software designer do
not take advantage of the Giants as much as they could.

It’s sad that the act
of building on the works of others, a necessity for progress, is labeled as
stealing by those who are blinded by dreams of easy money so much that they fail
to recognize that everyone, including themselves, is better off in a world where
we can freely build on ideas of others. Unfortunately those people have enough
money, bad taste and political savvy to force upon us copyright terms of insane
lengths and damaging patent policies.

Wise people recognized long ago that we all owe a big debt to those who came
before us for their ideas and whatever our contribution to the pool of knowledge
might be, it’ll be miniscule compared to collective knowledge gathered so far
that we were able to build upon.

If I have seen further it is by standing on the shoulders of Giants.
- Isaac Newton

It’s scary to think what will happen if patent system will degenerate to the point when
it’ll allow monopolizing fairly obvious yet crucial ideas. What would happen if
someone secured a patent for text editing system, making it impossible to create
a competing text editor. What would happen if someone secured a patent for
client-server communication making it impossible to create web (or one of
hundreds useful applications)? Everybody would suffer because the whole computer
industry would freeze. One lucky soul could charge monopolistic prices for the
only implementation of text editing idea but it would lack the incentive to
improve it. Both high-prices and lack of features would lead to meager adoption.
The vendor might think that he’s in heaven but everyone else would be much worse
off than in an alternative universe where thriving competition drives the prices
down and quickly improves the software making it useful and affordable for more
people, indirectly stimulating hardware purchases which in turn stimulates
creating more kinds of software. Ironically, the vendor would also end up with
smaller revenues than in the alternative universe because free competition would
create a market much bigger and one lucky winner that would dominate this market
would earn profits unheard of in any other kind of enterprise.

So let’s take
advantage of the fact that we’re not yet in a world like that. Thousand flowers
can still bloom and your mission is to take existing flowers and make them
slightly better. Let’s rip off each other’s ideas, let’s build on each other’s
work, let’s compete - that’s the way to make users happy. I claim that we don’t
do it nearly enough. You would think that because of fierce competition good
ideas in software would immediately get replicated but my observations rather lead me
to nod with understanding when I’m reminded of those words:

Don’t worry about people stealing your ideas. If your ideas are any
good, you’ll have to ram them down people’s throats
- Howard Aiken

Below are a few examples of neat software ideas that I’ve seen in at least once
and would like everyone else to copy in their products.

1. Emacs has a great
implementation of the "intelligent auto-completion" feature. While it’s a
standard feature in many (although not all) IDEs those days it usually only
works for a specific kinds of text e.g. it might auto-complete C/Java
function/method names and parameters and it requires pre-population of the
auto-completion data. Emacs will build the list of suggestions by parsing the
current text so it learns from what you’ve already typed. Works wonders - I
became depended on this feature a few hours after discovering it. Now every
editor that doesn’t have it is painful to use. Please, o unknown programmer,
steal this idea.

2. There are 2 popular ways to implement spell-checking:

• user has to invoke a menu item, gets a dialog box and is forced to
  sequentially go through every misspelled word in the document and decide
  whether it should be replaced, ignored, added to dictionary etc. Gets annoying
  if you spell-check the same  part multiple times and have to ignore the
  same words over and over again.
• pointing out misspelled words directly on the screen by e.g. putting read
  squiggly lines underneath and letting user to correct spelling using
  mouse-activated context-menus

The second method has been with us for quite some time and is obviously
better that the first one yet there are still programs that do it the hard-way
(e.g. Macromedia Dreamweaver).

3. There are more than 2 ways to implement searching in a text editor. One
popular way is to have a permanent Find dialog box with a field to enter text
we’re searching for and a button for Find Next. The dialog box will sometimes
obscure the text we’re working on requiring us to move it around. Some pervert
designers also force us to use mouse in order to switch the search directions
(i.e. either forward or backward) by making this option a drop-box or a radio
button. Why this design is still in use while there is a superior solution:
pop-up the dialog only for entering the text to search, pressing enter dismisses
the dialog and navigates to first match. We also have key strokes for Find
Previous and Find Next actions (I like to map them to F3/F4 keys). That way
there’s no need to use the mouse, no need to take your hand away from the
keyboard - that’s important for all the touch-typists out there.

Additionally Emacs has a small improvement of this idiom: it highlights all
the matches in the currently visible portion of the document.

4. Bookmarks in a text editor are quite useful when writing code. For example
imagine that you just wrote the .h file with function declarations and you’re
writing the implementation in a corresponding .c file. I usually find myself
unable to remember the exact details of what is the name of the parameter x or
what exactly is the name of a function etc. so I have to frequently switch
back-and-forth between the files to refresh my memory. If that’s only two files,
you can usually get by with other methods (just switching between buffers in
Emacs or arranging the display to see both files at the same time) but I found
that if the feature is well implemented then I prefer to use it even in two
files case and the more places you need to bookmark the more valuable it
becomes. But here’s the tricky part: "well implemented". It’s crucial that
navigating to bookmarked places is quick. I’ve seen only one editor that
implements it well: Source Insight. With a single keystroke (Ctrl-M) you can
both set a bookmark and navigate to a bookmarked place (don’t just believe me,
download free evaluation version of Source Insight and play with it). Sadly,
Emacs implements bookmarks poorly - key sequences are too long that managing
bookmarks is more trouble than it’s worth it. So if you implement a text editor,
here’s an idea for you: steal Source Insight’s bookmarks management. It’s
excellent.

5. People who use third party file managers (instead of built-in Explorer)
are usually power users. They might need to drop into command line to do things
that are not easy to do from GUI like grep through files, launch compilation
etc. I first saw this feature in Total Commander: there’s additional text edit
field at the bottom of the window. If you type something there and press enter,
it’ll be executed as a command in the directory that is shown in the active
pane. It’s a great feature that saved me a lot of time (as compared to
explicitly launching cmd window and mindlessly navigating to the directory that
I already have selected in my file manager). Yet not every file manager has it.
The result is that I won’t be using those file managers: once I’ve tasted the
usefulness of it, I refuse to use a tool that doesn’t give it to me.

Maybe we should have a collection of such "best design for doing foo"
somewhere on the web. Then if, for example, someone created a new text editor
and it lacked some of the useful features or implemented them in a way inferior
to existing implementation, we could send him a link to this page and say: dude,
we’re not gonna use your software until you implement all those features.
Please, do implement all the great ideas you have that make your text editor
better than others, but this is your baseline. Anything worse than that is not
gonna cut it.