On Fri, Mar 5, 2010 at 4:48 AM, iquiw <iku.iwasa@gmail.com> wrote:
Hi Johan,
On Fri, Mar 5, 2010 at 6:18 AM, Johan Tibell <
johan.tibell@gmail.com> wrote:
> Here's a proposal for a project I'd be willing to mentor:
> = A high-performance HTML combinator library using Data.Text =
Perhaps it's not scope of the project, but if compatibility doesn't
matter, I want new HTML library have uniform naming convention
for functions that based on element or attribute.
For example, function name should be;
- "e_" + element name ("html", "head", "body" => "e_html", "e_head", "e_body")
"a_" + attribute name ("href", "id", "class" => "a_href", "a_id", "a_class")
or
- "e" + capitalized element name ("html", "head", "body" => "eHtml",
"eHead", "eBody")
"a" + capitalized attribute name ("href", "id", "class" => "aHref",
"aId", "aClass")
or some other convention.
I think I would use the module system for namespacing rather than using function prefixes. Like so:
import Text.Html as E
import qualified Text.Html.Attribute as A
E.html ! [A.class_ "my-class"] (... more combinators ...)
(Assuming that "!" is used to introduce attributes.)
This allows you to use the element names and/or the attribute names unclassified if you so desire.
html ! [class_ "my-class"] (... more combinators ...)
Function names in the 'html' library are unpredictable from
corresponding element/attribute names...
("head", "base", "a" => "header", "thebase", "anchor")
I'm of the same opinion. The combinators should match the element/attribute names as far as possible. The rule that I had in mind was that the combinators should have exactly the same name as the corresponding element/tag except when the name collides with a keyword (e.g. "class"). If the name collides with a keyword we could e.g. always append a "_".
Cheers,
Johan