Book Review of “The ThoughtWorks Anthology: Essays on Software Technology and Innovation”

I would not be as quick to brand isBlank() as an example of bad code. It’s a static utility method and it is only required to do what it is asked for, but with a maximum possible efficiency. Indeed, the for-loop would look a bit ugly to a stranger, but strangers don’t look inside the code, they would just use it, unless they have the urge to criticize or a sudden curiosity attack.

See, your suggested “substitute” calls String’s trim, which (a) has TWO loops (to check from the beginning and the end of the String; (b) considers any symbol with a code lass than ‘ ‘ to be a whitespace, which is not what isBlank() doing (it is more strict and sticks to Unicode definition of whitespace). Not only that, a call to your string.trim() results in creation of a new String object, which you’re almost certainly not going to use (unless it _is_ blank and it gets cached by VM in the String pool). Think again: isBlank() does exactly what it was asked for, is more efficient and you don’t have to think about all of the above, unless you’re that wonderful essay author or just a reader with brain somewhat disturbed by essay’s content.

As I’ve wrote (which has other performance and memory characteristics).

“[...] results in creation of a new String object, [...]”

I know, see above.

“[...] , is more efficient [...]”

I know, see above.

To add to what vr said, another thing to think about is that Commons Lang isblank() is a library routine which is probably used a lot; I use it all the time. Its code is clean, simple, and straightforward so it will be easy to maintain. It may not be “pretty” from the perspective of a Ruby programmer, but like vr says, it’s maximally efficient.

Speed is still a significant issue with Ruby, perhaps because there is too much “pretty” Ruby code that’s inefficient.

Yes, as I’ve wrote:

“(which has other performance and memory characteristics).”

trim() probably does not create a new String in Java, because Strings are immutable (I know Reflection yaddayadda).

“trim() probably does not create a new String in Java” is that a typo? Since Strings are immutable I’d think that the obvious conclusion is that it would have to create a new String.

And if you look at the source, trim() calls substring(), which ends with

[pre]
return ((beginIndex == 0) && (endIndex == count)) ? this :
new String(offset beginIndex, endIndex – beginIndex, value);
[/pre]

Thanks for looking, it does create a new String. What I meant was that it would create a new String object but do not copy and use a new character array but reuse the old one, just with different pointers to beginning and end. Because String is immutable this would work.

I took a look at new String(), they do a array copy, perhaps because otherwise a larger String from which a substring was created could not be reclaimed by GC.

If you are using Java 6, there is a new method: String.isEmpty(). You could use that in your Java example and then you wouldn’t need to write your own.

basic pooker instructions…

shielded:gasser circumvented!purports:overwriting!esteemed:…

OMG!

public boolean isEmpty(String value) {
return value == null || value.trim().length() == 0;
}

(Or use isEmpty if ur string is not null and u under java 6).

That last “isEmpty” (OMG indeed) has the problem of creating an extra String object unnecessarily if the string is already trimmed.

Better:

public boolean isEmpty(String val)
{
return val == null || val.length() == 0 || val.trim().length() == 0;
}