<configuration>

……
    <system.net>
        <mailSettings>
            <smtp deliveryMethod=”Network” from=”name &lt;yourname@gmail.com&gt;”>
                <network enableSsl=”true” host=”smtp.gmail.com” password=”yourpassword” userName=”username@gmail.com” port=”587″/>
            </smtp>
        </mailSettings>
    </system.net>
</configuration>

So the enableSsl=”true” allows the SSL connection and you need to write out explicitly the right port number, 587 in this case (I tried 465 or leave it blank. In both cases, the smtp doesn’t work properly.).

A quick reminder: you can always set the deliveryMethod = “SpecifiedPickupPlace” to send email to local folder during development time

 今天你菊子曰了么？

一下的介绍基于上面的文章，主要是自己做个总结，方便自己找。

windows上远程调用linux端GUI程序的原理是在windows端启用一个x server, 然后将linux端的程序的显示forward给x server。由于linux管理显示及输入的都是通过一个统一的Server，所以本地和远程显示只是Server的地址不同，本质上并没有区别。

下载安装xming之后，启动。在windows 任务栏（时间显示旁边）可以看到一个X样的图标，说明server已经启动了。鼠标移上去可以看到 xming Server:0，记住这个数字。

启动putty，在左侧catagory: 选择Connection –> SSH –> X11，选上Enable X11 forwarding，在X display location里输入: localhost:0.

然后回到Session，建立连接就可以了。

这样做之后感觉反应速度比cygwin快，而且用鼠标右键就可以复制了。。。（老笔记本终于可以往窗口里复制粘贴了）

=================    疲倦的分割线   ==================

将正常小球记作N，不正常的记作A。

引理1：已知一个小球是N，A在两个小球之中，只需要一次就可以称出A。方法如下：

两个待选小球中任取一个和N比较，若平衡，则剩下那个是A，否则取出那个是A。

那么以下方法可以3步称出小球：

1. 取两组R, L 各四个小球称。不失一般性，标记为 R{1,2,3,4}; L{5,6,7,8};

若R和L一样，记作R=L，则剩下的集合S{9,10,11,12}中含有A，R，L构成标准组；取{9,10}与{1,2}比较，若相等，则A属于{11,12}，否则属于{9,10}，由引理，一步可以称出。

若R != L，假设 R>L (R比L重)。**交换{1, 2, 3} <–> {5, 6, 7}，若平衡不变，即仍有R>L，则A属于{4,8} (问题一步可以解决)。否则A属于{1,2,3,5,6,7}。其实这里故意绕了一下，这样的话是解不出来的，我们其实可以从这一步利用S = {9，10,11,12}是标准组得出更多的结论(知道A属于{1,2,3}或者{5,6,7}并且A是较重或者较轻)。

从**处开始：交换{1,2,3} <–> {5, 6, 7}，并用{9,10,11}代替{1,2,3}，此时R’={5,6,7,4}; L’={9,10,11,8};S’={1,2,3,12}

若仍有R’>L’，则A属于{4,8} (问题一步可以解决)。

若R’ = L’ 则A属于替换下天平的三个球，即{1,2,3}并且A较重（因为A来自于R)。{5,6,7}中任选两个比较即可。

若R’ < L’ 则A属于{5,6,7}且A较轻（因为A来自于L）。{1,2,3}中任选两个比较即可。

技巧的关键就是，通过交换天平两边的部分小球来进一步缩小A的范围，同时用标准组的球进行替换，设立比较的标准可以制造更多的不同结果（可能出现的结果越多，对于根据结果分类帮助更大）。

不过我不知道解法是否唯一。

大胆的进一步推测，因为第一步称只能产生2个结果，第二步通过轮换产生3个结果，第二步之后知道轻重，用2分法就可以比较快的得到答案。所以可以猜测，当总的小球数介于 [ 2*3*2^n, 2*3*2^(n+1) ) 时，用 (n+2)步就可以称出。不过我还没想到好的证明方法。

 本文用菊子曰发布

——————————————

Select / change data

1. select data:

select [xxx] from [yyy]

distinct col: select distinct [xxx], [zzz] from: get rid of duplicated rows

top n / top n percent (only show first several results from the select statement): n percent will round to integer using a ceiling rule, i.e. 6*10% = 1.

set rowcount n (0, default value — all rows)

as key word — aliasing (as is optional)

“+“: combine two strings

order by [xxx] desc/ asc

where clause (logical order: not > and > or)

where [xxx] like (wild cards)

functions on string: lower, upper, etc. 

select … into [tmp_table] from …

2. insert / update / delete data:

(*reminder: primary key doesn’t allow null values)

(1) insert (into) [table] (cols) values (vals1), (vals2) 

(*note: values shoule be consistent with each column)

(2) update [table] set [column] = xx

where [col] = / between / not in / in …

(*note: between includes two ends)

(3) delete from [table] where…

3. transactions: (explicite / implicite)

batch different statement in one transaction and if one of them fails, all the changes get rolled back.

begin tran

[statements]

…

commit tran

use of rollback / cube

———————————————–

View is a set of pre-defined SQL statements. It provides the interface between the user and database so that the user only has limited accessiblity to certain columns and a view can combine different columns from different tables and the user doesn’t need to know the underline structure of these tables.

1. create / drop a view, (alias, filter, Order by and top(100) percent etc.)

create view xxx

as

[statements]

drop view xxx

2. create index on a view (PK (primary key) is an index by default)

3. view encryption: hide the definition of a view (remember to backup your original definition other wise you can’t modify the view after encryption).

create view xxx with encryption

4. schema binding: you will get an warning if you are trying to change the underline table structure (it just gives out waring but not halts the process).

create view xxx with schema binding

5. update a view, if the base table is only one single table. [with check option], with this option, the insertion which doesn’t match view’s filter (like where) will not be executed.

create view xxx

as …

with check option

6. grant permission on views

grant select on xxx to guest

revoke select on xxx from guest

————————————————————————————-

For the following, I am just listing some points of SQL with some notes. This may actually be easier for review purpose:

1. Tables: works as the element in RDMS (relational database management system). Each table contains one or more columns. We could build relationships between different columns within one table or between different through foreign key constraint.

2. Data type: as in every language, MS SQL contains several data types:

* char, nchar, varchar, nvarchar. The prefix n- corresponds to unicode, takes double spaces for each characters ( 2 bytes comparing to 1 bytes). Each type here takes size as its parameter, like char(10) (default char() is one character). var- means the length can change with actual content.

* int, bigint, smallint, tinyint, decimal/numeric, float, real, money, smallmoney: int (-2^31 ~ 2^31-1), bigint(-2^63~2^63-1), smallint( -2^15, 2^15-1), tinyint(0~255), decimal(precision(-38, 38), scale), float()/real-float with given number of digits. money (-922 tri – 922 tri) without currency info. smallmoney(-200000, 200000)

* date, datetime, datetime2, smalldatetime, datetimeoffset, time: datetime2 (to 0.1 ms) has higher precision than datetime

* geometry, geography:

* uniqueidentifier, rowversion:

* binary, varbinary, bit, image (binary info.): var- variable size of binary data.

* xml:

3. Use Microsoft Visual Studio / DBMS to create database (?? what’s the use of identity specification ??)

4. Add foreign key relationship through table designer (the parent column needs to be primary key) — the relationship can be viewed graphically from database diagrams

5. calculated column: we can add calculated column by:

alter table xxx add “calculated column” as [formula]

drop a column:

alter table xxx drop column “column”

we can create the calculated column as persisted, in which case this column actually takes one space on hard drive and when queried, the data is grabbed from there directly without real calculation. This is useful when the data is deterministic and saves you time:

alter table xxx add “persisted column” as [formula] persisted

一般来说有两个版本可以用，一个是express版，说是适用于学习和小型数据库的管理操作。另一个是enterprise evaluation，包括所有的功能，不过只能免费试用180天，所以。。。也是用来学习的。

今天就只说说安装。如果从来没有安装过，那就选全新安装(stand-alone installation)。期间会检查各种你应该装没有装的东西。忽略所有的warning，只看failed的项目，然后如实装上就可以了。我自己又不得不装了microsoft power shell 和 visual studio 2008 sp1。

重点来了：折腾完这些，就在Server configuration (服务器配置) 那里卡住了，怎么弄都说我的用户名密码invalid云云。在google上转了半天，发现这个错误主要来自两个原因（都和错误信息没有什么关系）：

1. 密码需要设置为强密码（Strong password）。多强叫强？我也不知道。从结果看用数字和字母（我只用了小写字母）就通过了~。

2. 用户名的问题。我是用默认的administrative的账号登陆的xp，然后至少我没有办法把这个用户作为Server client的用户，其他随便起的名字也不行。一个解决办法是打开Control panel，新建一个用户，选择limited access （不知道administrative行不行）。然后在MS SQL的Server configuration里面选择所有账户使用相同用户名密码，之后选择浏览 （browser），在分组和computer用户下面的框里搜索自己之前建立的用户（应该可以找到结果），添加进去就可以了。

以后一路无险，应该算是安装完成了。。。