Rfcs: Adopt a policy of 1 sentance per line for text

Created on 7 Dec 2016  ·  10Comments  ·  Source: rust-lang/rfcs

That way, diffs between commits have a fighting chance of being readable. Realistically, I think everyone reads these on GitHub or at least has access to a markdown renderer, so line width is unimportant for readability of the documents themselves.

T-core

Most helpful comment

http://rhodesmill.org/brandon/2012/one-sentence-per-line/ calls this "semantic linefeeds". The point is not to keep entire sentences unbroken, it is to avoid re-wrapping an entire paragraph when editing it (e.g. inserting a few words in the middle). To do that, break not to maximize the length of each line within the set maximum width, but break at "logical" points like between sentences, after commas, etc.

An example from this article:

 ...
 the definition in place of it.

-The beauteous scheme is that now, if you change
-your mind about what a paragraph should look
-like, you can change the formatted output merely
-by changing the definition of ‘‘.PP’’ and
-re-running the formatter.
+The beauty of this scheme is that now, if you
+change your mind about what a paragraph should
+look like, you can change the formatted output
+merely by changing the definition of ‘‘.PP’’
+and re-running the formatter.

 As a rule of thumb, for all but the most
 ...
 ...
 the definition in place of it.

-The beauteous scheme is that now,
+The beauty of this scheme is that now,
 if you change your mind
 about what a paragraph should look like,
 you can change the formatted output
 merely by changing
 the definition of ‘‘.PP’’
 and re-running the formatter.

 As a rule of thumb, for all but the most
 ...

All 10 comments

so line width is unimportant for readability of the documents themselves.

Of course it's important! It's important to the person writing it. Everything I write in my editor is wrapped to 79 columns with appropriate line breaks.

It's 2016 you know, and editors can display text wrapped even if the underlying text isn't :).

http://rhodesmill.org/brandon/2012/one-sentence-per-line/ calls this "semantic linefeeds". The point is not to keep entire sentences unbroken, it is to avoid re-wrapping an entire paragraph when editing it (e.g. inserting a few words in the middle). To do that, break not to maximize the length of each line within the set maximum width, but break at "logical" points like between sentences, after commas, etc.

An example from this article:

 ...
 the definition in place of it.

-The beauteous scheme is that now, if you change
-your mind about what a paragraph should look
-like, you can change the formatted output merely
-by changing the definition of ‘‘.PP’’ and
-re-running the formatter.
+The beauty of this scheme is that now, if you
+change your mind about what a paragraph should
+look like, you can change the formatted output
+merely by changing the definition of ‘‘.PP’’
+and re-running the formatter.

 As a rule of thumb, for all but the most
 ...
 ...
 the definition in place of it.

-The beauteous scheme is that now,
+The beauty of this scheme is that now,
 if you change your mind
 about what a paragraph should look like,
 you can change the formatted output
 merely by changing
 the definition of ‘‘.PP’’
 and re-running the formatter.

 As a rule of thumb, for all but the most
 ...

I like to read the raw (on GH, this allows commenting at the same time as reading) rather than rendered text and this would make it much harder, so -1 from me.

@aturon Ping; Is this something we might want to do? (Personally I think formatting can be left up to the person writing the RFC...)

@Centril I don't think this is worth the hassle.

@aturon I agree. Shall we (you) close / fcp close then?

Since Aaron didn't think it was worth the hassle; I'll just close this unilaterally to reduce the backlog.

If you think this was done in error, please speak up :)

FWIW I like semantic linefeeds and use them in my own writing (including RFCs), but I agree that it’s not worth making everyone to reformat their submissions.

Was this page helpful?
0 / 5 - 0 ratings

Related issues

onelson picture onelson  ·  3Comments

Diggsey picture Diggsey  ·  3Comments

3442853561 picture 3442853561  ·  3Comments

torkleyy picture torkleyy  ·  3Comments

mqudsi picture mqudsi  ·  3Comments