[erlang-questions] Style wars: junk comments

Pierpaolo Bernardi <>
Wed Sep 12 13:19:29 CEST 2012


On Wed, Sep 12, 2012 at 12:27 PM,  <> wrote:
>> On Wed, Sep 12, 2012 at 11:12 AM, Ivan Uemlianin <> wrote:
>>> If you are reading code on paper, it helps if functions are positioned
>>> predictably in the stack of paper.  Alphabetic ordering of functions and
>>> grouping functions into sections are two good ways of doing this.
>>
>> "Reading code on paper" has already been mentioned some times in this
>> thread.
>> But really people read code on paper, nowadays?
>
> Yes, me, lots.
>
> Let me offer you a paradox, half fun and full earnest.'
>
> Fancy text editors and IDEs are tools for NOT reading code.
>
> If you actually want to read code, paper (ideally augmented
> with literate-programming-style automatically generated
> indices) is way better.  If you don't know what I'm talking
> about, look at Knuth's "The Stanford GraphBase".

I have this book. Love the content. Hate how it's presented.

(And I wish Knuth had abstained from using the ignoble, hackish,
sometimes not complying to C of any standard, tricks, which makes the
code presented hard to reuse, and instead had concentrated on a higher
level view.)

> Just today, I found a bug in a Smalltalk class that had
> defeated me for a day because using an IDE is too much
> like tunnel vision.  Given a printed listing, it took me
> half an hour to read the whole class, and the bug practically
> jumped up and down and shouted to get my attention.

I have little experiences with Smalltalk, but when I tried it, it's
usual IDEs gave me the impression of looking at the code through a
keyhole.
So I understand that in this case printed source is liberating.

P.



More information about the erlang-questions mailing list