[TYPO3-dev] Are performance improvements part of maintaince releases?

[JoH asenau]

> or in other words:
> is a function that makes 50 identical sql requests when rendering a
> page a bug?

It depends on the personal definition of "bug" I guess.

If you just look at it from the theoretical point of view of a computer
scientist, it is no real "bug", as long as the function passes all unit
tests. - By this definition a bucket wheel excavator will be an appropriate
vehicle for public transport because it can be used to move people from A to
B.

The problem in the definition is that the word "appropriate" is missing. To
stay with the bucket wheel excavator example: Of course you can use a few
hundred people with shovels or maybe a few thousand people with spoons to
dig out the same amount of coal but the excavator will do it in a much more
efficient way.

If you look at it from an economical point of view, it is obviously a bug to
send 50 identical requests, because it wastes precious resources that could
have been used otherwise and it will even have ecological impacts, which
might have become even more important in the 21st century.

The problem is, that most of the purists among computer scientists do not
accept a bug in the concept of their code as a "real" bug, as long as the
behaviour of the code is the one they expected. IMHO this is one of the
major problems we have to face when looking under the hood of TYPO3: There
are many of those conceptual bugs - I guess even more than "real" bugs -
that decrease the performance and the usability a lot.

I could accept the argument, that maintenance releases are not about
performance tuning, because this is up to major releases. But when you take
a look at 4.3, this is currently much _slower_ than its predecessor, even
though some computer scientists introduced the so called "caching
framework". Why is it slower? Because the caching framework is just another
conceptual bug - that still passed the unit tests. SCNR ;-)

So IMHO it is crucial to remove these conceptual bugs whenever, wherever and
however we can. Even within maintenance releases, because everybody should
be able to benefit from these improvements.

As long as you can improve a piece of code "under the hood" regarding speed
or consumption of resources but without changing the expected behaviour, it
should be considered to be buggy - these bugs *must* be fixed and fixing
them is definitely no "new feature"!

Just my 2 cents

Joey

-- 
Wenn man keine Ahnung hat: Einfach mal Fresse halten!
(If you have no clues: simply shut your gob sometimes!)
Dieter Nuhr, German comedian
Twitter: http://twitter.com/bunnyfield
Xing: http://contact.cybercraft.de
T3 cookbook (2nd edition): http://www.4any1.de
TYPO3 Schulung: http://workshops.eqony.com