||
This is an in-mail from TYUST.
本期开始加开窗口,推出科学网特色博主,有用链接等。
心中有块石头...
(接上回*)定理1.6证明(第二段后半部分...其实是第三段) 。
Let C=1/2A. Then we have
lct(X, B, |A|R) = 1/2 lct(X, B, |C|R),
hence it is enough to give a positive lower bound for lct(X, B, |C|R).
评论:取一半A究竟起到何种本质作用?
Let n, m, eps' be the numbers given by Proposition 5.9 for the data d, r, eps.
评论:n, m, eps' 的存在?(迄今达到Pro5.9的条件了吗?)
Pick L∈|C|R. Let s be the largest number such that (X, B+sL) is eps'-lc.
评论:s一定存在吗?
It is enough to give a positive lower bound for s. In particular, we can assume s<=1.
评论:上界有正的下界,则上确界有正的下界?(须温习上界、下界方面的基本知识、结果)。
.
小结:完成第三段(疑问不少)。
.
加把劲,进入第四段(原文占7行)。
There is a prime divisor T on birational models of X such that a(T, X, B+sL) = eps'.
评论:这是向命题5.9靠拢,但明显存在“隐含的调用”。(泛函方程?)。
Let x be the generic point of the centre of T on X. Assume x is not a closed point.
评论:继续向命题5.9靠拢。
Then cutting by general elements of |A| and applying induction, there is a positive number v bounded from below away from zero such that (X, B+vL) is lc near x.
评论:看不出(?)。
Then (X, B + (1-eps'/eps)vL) is eps'-lc near x, by Lemma 2.3, because
B + (1-eps'/eps)vL = eps'/eps B + (1-eps'/eps)(B + vL).
评论:这一步该能看懂(引理2.3是个泛函恒等式)。
In particular, s>=(1-eps'/eps)v. Thus we can assume x is a closed point.
评论:前一句是根据s最大。后一句看不出(?)。
.
小结:定理1.6证明部分,第四段完毕。(还有两段,共9行...胜利在望啊)。
Archiver|手机版|科学网 ( 京ICP备07017567号-12 )
GMT+8, 2024-11-23 19:11
Powered by ScienceNet.cn
Copyright © 2007- 中国科学报社