Hello Wren,
On Sun, Sep 21, 2008 at 10:02 PM, wren ng thornton
By intention, probably. But by accident it is a lot more common than you might think. Accidents like corrupting the linebreaks or line wrapping in a file are quite prevalent when exchanging files across different platforms or medium formats. Consider also file type misdetection with other languages like HTML; or more convincingly, consider language interleaving. And there're always garden variety typos like dropping a bird track or adding a space before it.
Usually such corruption is easily captured by code not compiling, but I think the big thing to consider when suggesting changes to the .lhs format is how it affects the robustness of detecting file corruption, typos, etc. The whitespace restriction captures a goodly number of typos mentioned above, though it may not be so helpful for the other corruptions.
Thanks for your thoughts. I see your point and I agree it's an important one. For example, as you well know, if you produce HTML with our Haskell libraries, there's some good chance you'll end up with a ">" mark in the beginning of the line, because of the way the tags are written. However, that's HTML. If you decide to interleave with other languages, I don't think the probability of having a ">" unexpectedly showing up in the beginning of the line is enough to justify making this a rule. In most languages it would be like writing a condition very close to the end of the line and having it break in the middle: if (x
y)
I think the restriction made sense when the report was written, but what I'm arguing is that given new developments we have made, the benefits of lifting the restriction are greater than keeping it. At least for me! : ) To be sure, I'm weighting the chances that the mandatory blank lines will save you versus the possibility of writing in Bird-style more freely. Again, please do point out other flaws in my argument. Thanks, Paulo