Don't use the MAX1232 here. It's not well suited for micros needing an active low going reset.

Just a small question to Kaai. What do you say that MAX1232 doesn't work well with active-low reset chips? An active low is an active low isn't it?

The reason is that the MAX1232 does not work down to Vcc=0V, but only about Vcc=1.5...2V. I use a trick to make it work down to Vcc=0V by using a pull-up of 2.35kOhm, like shown in this example (the jumper is only opened during ISP- programming):

...

In combination with the internal pull-down of 50...300kOhm of AT89S52, this assures proper logic high level even when Vcc is such low, that the P-MOSFET at output of MAX1232 cannot be turned on, but looks like a very high impedance to Vcc. When the supply voltage of MAX1232 is too low, then the output behaves like a tristate. Here the pull-up helps to force logic high level during the entire reset.

With an active low reset micro situation looks completely different, because the /RST output of MAX1232 is only open drain and not push-pull like the RST output! To assure proper performance of reset down to Vcc=0V, like above, would mean to use a pull-down resistor at /RST output. But this will only work with a push-pull output, not an open drain!!