Andrew's ramblings

Couldn't agree more, I think it's good practise. It gets you into the habit for when you are writing XML and also HTML/XHTML. Drives me insane when I read code that has open tags everywhere!

Close close close! And code your CF Modules accordingly.

It makes code so much more readable and (until HTML5 is properly out anyway) it sets good muscular memory for marking up in XHTML

It'd be good if you *could* clarify the point I mention above, because it's the basis for your rationale, and it pretty much sounds like ballocks to me.

There are situations in which one could argue closing a <cfinvoke> tag makes sense, because it can have an optional closing tag. And there are a few other tags like that. But MOST CF tags are not like that, so there is very little scope for ambiguity. There is never any doubt, for example, as to whether a <cfquery>; or a <cfqueryparamt> tag are the beginning of a block of code or not ("yes, always" and "no, never", respectively).

So, sure, close tags which have some ambiguity about them, but you make no case for closing *all* tags.

Because, basically, there's no *good* case to be had for doing it.

--
Adam

FIrstly, completely fair enough re custom tags.&nbsp; The ambiguity is definitely possible there.

However you seem to be missing the point that in XML and XHTML it's a *requirement of the specification* that all tags must be closed.

This is *not* the case in CF. It's *absolutely* not the case with CF.&nbsp; Not all tag based implementations have that rule (HTML is an obvious example). Most CF tags DO NOT HAVE a closing tag. Try this:
<cfset foo="bar"></cfset>

(don't forget that <sometag /> is an abbreviation of <sometag></sometag>)

You get a compile error. Why? Because <cfset> doesn't have a closing tag. It's not that it doesn't do anything, it's that IT DOESN'T HAVE ONE.

What you're doing when you do this: <cfset foo="bar" /> is not "closing the tag", all you're doing is putting an unnecessary slash in your code.

Also consider what the CF compiler does when it sees <cfset foo="bar"> vs <cfset foo="bar"/>. Nothing. It just ignores it. It does not go and process the "closetag" part of <cfset>. Why? Because there isn't one. Because <cfset> doesn't have a closing tag.

So closing a CF tag which has no close-tag implementation is:

* generally no clearer in intent (other than exceptions like <cfinvoke>, in which case common sense dictates one could possibly close the tag);

* actually - stictly speaking it obscures intent, because it suggests in other situations there MIGHT be a closing tag to look out for (in other people's code who don't slavishly, and IMO moronically close their tags);

* serves no purpose;

* is actually illegitimate as far as the language goes (ie: there very much is no </cfset> tag);

* based on a false premise that there's any justification to do so because tags intrinsically are meant to be closed.

Some people have said "oh, I just got in the habit when I was doing all that XHTML". Oh *please*. Are you really so incapable that you can't remember when you're doing XHTML and when you're doing CFML? If anyone ever admitted that to me, I'd be questioning whether they're in the right job (actually: no question necessary: they're not in the right job).

This leaves "following the Adobe coding standard". I think the Adobe coding standard needs to be changed, because it's currently flawed. I think one should definitely follow a coding standard, but one definitely should not follow that particular one until it's corrected.

--
Adam