虽然我远没有看完Paul Richard Halmos（1916~2006）的所有数学教材，但我确实看了几本话题比较大众化的——因而同类教材也比较多的——从而在比较的基础上可以支撑我对Halmos的偏爱。所以我仍然能够直接说，我特别喜欢Halmos的数学书。
我看过他写的集合论、测度论、有限维向量空间和希尔伯特空间。这些书除了测度论之外都是很薄的。
至少我还不完整地看过他的自传——I want to be a mathematician。这个自传的开头就先明确地反醒了作者的人格特质，读来发现与我本人十分贴近，这足以支持我偏爱他的著作。这段话对其他读者来说也可以说明Halmos很可能是总能写出好教材的数学家。
I like words more than numbers, and I always did.
Then why, you might well ask, am I a mathematician? I don’t know. I can see some of the reasons in the story of my life, and I can see that chance played a role at least as big as choice; I’ll try to tell about all that as we go along. I do know that I wasn’t always sure what I wanted to be.
The sentence I began with explains the way I feel about a lot of things, and how I got that way. It implies, for instance, or in any event I men for it to imply that in mathematics I like the conceptual more than the computational. To me the definition of a group is far clearer and more important and more beautiful than the Cauchy integral formula. Is it unfair to compare a concept with a fact? Very well, to me the infinite differentiability of a once differentiable complex function is far superior in beauty and depth to the celebrated Campbell–Baker–Hausdorff formula about non-commutative exponentiation.
The beginning sentence includes also the statement that I like to understand mathematics, and to clarify it for myself and for the world, more even than to discover it. The joy of suddenly learning a former secret and the joy of suddenly discovering a hitherto unknown truth are the same to me—both have the flash of enlightenment, the almost incredibly enhanced vision, and the ecstacy and euphoria of released tension. At the same time, discovering a new truth, similar in subjective pleasure to understanding an old one, is in one way quite different. The difference is the pride, the feeling of victory, the almost malicious satisfaction that comes from being first. “First” implies that someone is second; to want to be first is asking to be “graded on the curve”. I seem to be saying, almost, that clarifying old mathematics is more moral than finding new, and that’s obviously silly—but let me say instead that insight is better without an accompanying gloat than with. Or am I just saying that I am better at polishing than at hunting and I like more what I can do better?
P. Halmos (1985), I want to be a mathematician, Springer
de Gennes (1976)的Journal de Physique Lettre，就是这一例子。他认为弹性的逾渗跟电导率的逾渗在模型上是等价的，因此弹性逾渗的临界指数就是电导率逾渗的临界指数。冯奚乔的论文是最早打破这一认识的论文之一。这也是他的学位论文工作。今天我们知道弹性逾渗与其他输运性质的逾渗是不同的，但仍然有很多open questions。特别是从弹性（线性弹性体）或粘度（Navier–Stokes流体）推广至流变学上的一般认识：粘弹性后，能否仍然去讨论“粘弹性逾渗”与纯几何的连接性逾渗的差别？从研究论文来看，我们不幸地看到，非但在冯奚乔以及大量其他人的工作之后，许多不假思索的实验家仍然按de Gennes (1976)的认识去粗略解读自己的实验数据；就算粘弹性临界现象也已经被大量报道之后，大量不称职的实验家也仍满足于弹性逾渗与几何逾渗临界点重合的假定、de Gennes (1976)的假定、超标度假定……试图反过来去下关于凝胶结构的结论。
这种完全无视理论认识的最新进展，对“什么未知、什么已知”的无知，对“实验家的当前任务是什么”问题的轻视，不幸地占据了大部分论文版面，影响了一代又一代的年轻人，从而淹没了正确的研究主线。90年代之后的科学界，进入了庸俗的publish or perish狂躁派对。冯奚乔的工作，包括逾渗的工作，或者在理论物理上稍微严肃些、形式些的兴趣，都显得太枯燥、太难引起大量波澜和impact，从而使实验的投入和“产出”不成正比，被迫变成了“隐学”。现在你再提起，就会被人评价一句：你这个问题太老了，现在没人感兴趣。
