On 2008-02-19, Peter T. Daniels wrote:

>> > Aha. that's why your sauces are low-quality. Considering the further
>> > use of the roux, you want exact measurement (if you are still
>> > measuring) of both flour and butter separately, also because between
>> > 10 and 16% of your butter, by weight, has evaporated during the
>> > melting. Trying to translate into weight from volume is bad enough.
>> > Sticking to the "equal volume" nonsense where flour weight equal to
>> > butterfats weight (i.e. 50 to 60 by weight in the Escoffier measures)
>> > is required throws you even further off .

>> I've never done any measuring for a roux. I just chuck a knob of butter
>> into the saucepan, melt it and add a bit of cornflour (starch). If it's
>> still runny, I keep adding cornflour until it's the right consistency.
>> The same with the milk or stock.

> Only the cheapest Chinese takeouts [UK: takeaways] (still) use
> cornstarch for thickening. The result is gummily awful.

There's nothing inherently wrong with it. Unlike flour, cornstarch is
almost pure starch, which is what you really want to thicken a sauce.
(Other components of flour require simmering and skimming to clarify
the sauce.) And Escoffier approved:

Though he was a traditionalist in many ways, Escoffier openly looked
forward to the day when pure starch would replace flour as the
thickener in stock-based sauces.

(McGee _On Food and Cooking_, second edition, in the section on
"Sauces thickened with flour and starch".)

The wonderful thing about standards is that there are so
many of them to choose from. [Grace Murray Hopper]