First of all, the LZ's statement is definitely correct. Calculations can certainly be done without variable delays (even in a for loop). Of course, the "calculation" here is very vague, leading to everyone having their own understanding. The implied meaning here by the LZ is that call can be used instead of delay because call appears in his code. Secondly, the LZ's code is also correct, with correct syntax and achieving the desired effect. What I want to show is that the LZ's code cannot prove the implied meaning in his view because I can achieve the same purpose in a non-call and non-delay way, so there is no way to explain the significance of call replacing delay. -------------------------------------------------------- From the first reply of ppdos, it is known that he understands the default delay effect of set /a I mentioned in building 14, but just didn't connect it with the LZ's statement as a problem. Simplification itself is not a problem, and I just think that condensing into one line of code is a bit redundant. One can also find evidence to prove the view that call can replace delay. (for /f %i in ('dir /b') do set /a n+=1 >nul)&call echo There are %n% files(folders) under the directory can be replaced by (for /f %i in ('dir /b') do set /a n+=1 >nul)&echo There are!n! files(folders) under the directory. This shows that when encountering variable delay problems in the connected statements, call can be used to replace, while metoo looks into the for, of course, not seeing anything valuable. -------------------------------------------------------- As for the signature of Wunaaihe, that's really off-topic. I only remember those few lines of words now, and the code has long been forgotten completely, and I don't bother to dig it out again, and just let it be whether it's continuous assignment or not. [ Last edited by qzwqzw on 2008-6-25 at 10:28 PM ]