Why would you think that I have not read the thread? Considering that I personally posted a good chunk of it, that would seem to be a strange assumption. Perhaps you would like to specify what I said that is not true?
As for the silly statement that launched all of this, see Re: Re (tilly) 4: millions of records in a Hash. It is silly to think that having available a micro-optimization like knowing the exact size of your keys is enough to make it worthwhile to rewrite something a dbm in Perl - for performance reasons! Such comments indicate an utter lack of perspective on what dbms do for you, and how much overhead Perl adds. You may not understand why that is silly, but it is. TMTOWTDI is fine and dandy, but not all ways are created equal...
And about professionalism. I didn't bring up that topic, nor did I ever say that someone should act in a particular way because "that is what professionals do". Instead I gave specific reasons that cause professionals to act as they do. It is the reasons that should matter.
Finally I suspect you have prematurely jumped to conclusions about how specialized I am. We have under discussion a very specific problem (fast storage and retrieval in Perl of more key/value pairs than will fit in a built-in hash) and that problem has a specific signature. If we had a different problem with a different signature, I would hopefully treat it appropriately. |