今天踢得真臭

最后一个大脚偏了45度,球直接飞到国会山前面的大湖里去了

还好不深,光着脚丫进去把球捡出来了

好几脚没来得及出脚,球就被人抢走了

近距离传球直接踢出边线

我这水平也就倒数吧

//上礼拜六进了三个球,绝对是运气

end of fiscal year

the project is successfully delivered, and customer is happy, said "fantastic job"

the process is painful, and it’s stressful during the last two weeks

yesterday afternoon i resolved the last issue and feels relieved

many obstacles stands in the process

for this small project (a couple of million level), at least 5 parties involved (not counting customer):
1) our dev team, design, prototype, develop, test
2) data center team, prod infrastructure, deployment
3) ops support team, dev/qa infrastructure, dns, firewall rules, smtp
4) security team, ato, scan
5) fed manager
6) procurement team

about 2 to 3 meetings every week.

recipients number in email threads can easily go up to 10 or more

communication and response is slow, sometimes requests are ignored.

5 month is wasted due to initial wrong procurement and later re-procurement

security team jumped in, jumped out and jumped in again, due to confusion of ato.

everything has to be prepared and requested early, as it normally takes 1 ~ 2 week get it done. for instance, it takes about 2 weeks to get a dns

initial documentation process is also time consuming, review, review and review, each review may take 1 to 4 weeks.

many things need authorization and justification

it’s feeling bad and frustrated when many little stuff are not in control and have to rely on non-responsive others, such as helping customer install software, requesting vm, requesting dns, requesting smtp whitelist,

lots of restrictions due to security concerns. everything has to be certified.

and last ridiculous scheme: contract ends 9/29 and thus we cannot officially help customer unless being approved by fed boss. it’s like a big bang and then suddenly end. not healthy, not friendly.

comment

最近把两年前的代码拿出来,需要一些修改用到新的项目上
首先是得理解之前的代码

理解代码又要从理解变量开始

看到了一个变量是这样的

L = 2 * max_lags + 1

为什么是这样呢?半天没想明白。

又看代码,发现把一个channel的数据往前移了max_lags

后来再想想,终于明白当初为什么设定L = 2 * max_lags + 1

这么做其实很不方便,L dependend on max_lags

而当时的情况是这样的:无法确定哪个channel在前面。而算法本身要求某个channel必须在另外一个channel后面。所以就想了这么个办法,先把一个channel往后挪max_lags,然后把L订成2 * max_lags + 1,这样总能cover。

而现在情况不一样了,第二个channel肯定总是在第一个channel后面。需要做反方向移动。所以代码得重新改改。

因为这个教训,以后写代码得好好注释一些重要的决策,以及上下文

white house’s memo on Using Evidence and Innovation to Improve Government Performance

here: The President recently asked his Cabinet to carry out an aggressive management agenda for his second term that delivers a smarter, more innovative, and more accountable government for citizens. An important component ofthat effort is strengthening agencies’ abilities to continually improve program performance by applying existing evidence about what works, generating new knowledge, and using experimentation and innovation to test new approaches to program delivery. This is especially impmiant given current fiscal challenges, as our nation recovers from a deep recession and agencies face tough choices about how to meet increased demand for services in a constrained resource environment.

官僚

还是scan

上个礼拜五写信,不回,找人,不在

礼拜一提交ticket,不回,找人,不在

礼拜二上午找人,找到。他问什么时候要。我说越快越好。可以今天或者明天弄。

礼拜二过去,没消息

礼拜三,另外一个等我的team着急了,我就cc给那个哥们,也没回

礼拜三过去,还是没消息。。。。

头也不管

真是有点儿受够了。。。。