Friday, December 15, 2006

Use Numbers

To argue, try to use statements with numbers, like
  • "1000 sites * 20 services * 200 VOs * 10 roles/groups = 40M metadata"
  • "100k CPUs * 10 jobs per day * 10 queries per job = 10M queries per day, 115 queries per second."
  • Publishing one service adds 138 RDF statements to the triple store, which consumes up about 25 KB of memory. Then each triple uses 181 bytes. Each metadata has at least 3 triples: (entity, hasMetadata, metadata), (metadata, hasMetadataType, type), (metadata, hasMetadataValue, value). Thus 4M metadata uses 181 * 3 * 4M = 2.1GB.
And pay attention to detail.

Wednesday, December 06, 2006

Tuesday, December 05, 2006

Create A User Community to Improve Our Open Source Software

  • Create impact. Open source software's value relies on its usage.
  • Get feedback, which could be good ideas for the future work.
  • Create a well organized website, easy downloads, a bugzilla system, etc. Twiki and Sourceforge are very helpful in this sense.
  • Well answer questions by email.
  • Create a user list, which includes the project name, url, contacts, etc.
BTW, check out OpenCola, which applies the open source idea in making cola.

Monday, December 04, 2006

My New Laptop

Today I got a new laptop, a SONY VAIO, which has
              • Intel® Core™ 2 Duo T7200 Processor 2.0GHz
              • 2GB, 100GB HDD & DVD/RW Super Multi DL Drive
              • 13.3" WXGA TFT Display, Bluetooth & FREE T-Mobile WAN Card
              • ... ...
              • And a fingerprint reader! So that I am able to log in Windows by using my fingerprint. Cool!

Sunday, December 03, 2006




八阕】美国大学与雇主协会,每年按4个季度分别发布新毕业的大学生、 硕士和博士研究生的薪水调查报告,这报告共涉及80个学士专业,40个硕士专业,23个博士专业。冬、春、夏三季发布的数据是各季度的数据,秋季版发布的 是去年8月31日到今年9月1日为止的整个学年数据。以下列出的是该协会最新公布的美国博士和硕士毕业生2006年秋季薪水调查报告。


1、商业管理(Business Administration/Management):$81,438

2、电气/电子与通信工程(Electrical/Electronics & Communications Engineering): $81,297

3、冶金工程(含陶瓷科学与工程)(Metallurgical Engineering ):$80,667

4、计算机科学(Computer Sciences):$76,630

5、化学工程(Chemical Engineering):$75,659

6、计算机工程(Computer Engineering):$74,750

7、工业/制造工程(Industrial/Manufacturing Engineering):$73,292






3、社会科学(Social Sciences):$48,487

4、生物/生命科学(Biological Sciences/Life Sciences):$43,916

5、英语语言与文学(English Language & Literature/Letters ):$41,405


01、计算机科学(Computer Sciences):$71,165

02、石油工程(Petroleum Engineering):$68,833

03、电气/电子与通信工程(Electrical/Electronics & Communications Engineering):$66,687

04、计算机工程(Computer Engineering):$66,545

05、地质及相关科学(Geological & Related Sciences):$64,111

06、材料科学(Materials Sciences):$63,500

07、航天/航空/宇航工程(Aerospace/Aeronautical/Astronautical Engineering):$62,811

08、工业/制造工程(Industrial/Manufacturing Engineering ):$61,273

09、机械工程(Mechanical Engineering):$61,234

10、化学工程(Chemical Engineering):$59,008



最低年薪:$64,620,最高年薪: $143,620

在公司和企业管理层谋职(Management of companies and enterprises):$126,250

在联邦政府谋职(Federal Government):$108,090

在法律机构谋职(Legal services):$99,580

在地方政府谋职(Local government):$73,410

在州政府谋职(State government):$70,280



外科、普通外科(Surgery, general):$255,438


内科(Internal medicine):$155,530

小儿/青少年科(Pediatrics/Adolescent medicine):$152,690


家庭全科(Family Practice):$150,267



高中毕业平均年薪:  $28,645

大学毕业平均年薪:  $51,554




  • 选择专业前先要做好各个专业的benchmark, 什么专业工资最高。看来benchmark真是无时无地不重要。
  • 计算机科学博士起薪只比计算机科学硕士高7.7%。虽然理论上博士比硕士有更好的成长空间,究竟读不读博士还是要慎重考虑。
  • 医生收入压倒性的高,远高于其他专业,甚至包括律师。还是学医收入高。
  • 读书有用啊。读书越多,收入越高。
BTW, I read this from Ian Foster's blog:

"Indeed, it was outsiders—those with expertise at the periphery of a problem's field—who were most likely to find answers and do so quickly."

Friday, December 01, 2006

Java I/O

This presentation gives a good overview of Java I/O, from which I have learned:
  • FileInputStream/FileOutputStream: for sequential I/O and simple to buffer
  • RandomAccessFile: for full random-access
  • FileChannel: adds support for NIO byte buffers
  • MappedByteBuffer: for memory-mapped I/O
  • Buffering: for instance, use BufferedOutputStream for buffering FileOutputstream
  • Forcing I/O, e.g. flush() or FileChannel.force(), equals to the "fsync" system call
  • No direct I/O in Java