Training new Browser Engineers

From Genecats
Jump to navigationJump to search

parasol

Trainer: Galt Barber

parasol is software for running batches of jobs on the cluster of computing machines.

There is a full page of parasol information and examples here: Parasol_how_to

software engineering best practices

Trainer: Kate Rosenbloom

Recommended reading:

  • The Art of UNIX Programming , second part of Chapter 1 (Basics of the Unix Philosophy, Eric Raymond). pp. 11-27
  • Beautiful Code, Chapter 13 (Design of the Gene Sorter, Jim Kent). pp. 217-228


Some useful acronyms:

  • DRY programming (also OAOO) - Don't Repeat Yourself, Once & Only Once
  • YAGNI - You Ain't Gonna Need It

expressions:

  • Software rot, bit decay
  • Technical debt

and aphorisms:

  • Do the simplest thing that can possibly work (Ward Cunningham, Extreme Programming)
  • Any fool can write code that a computer can understand. Good programmers write code that humans can understand. (Martin Fowler, Refactoring)


Some philosophers of good practice:

  • Brian Kernighan, PJ Plauger, Rob Pike, Yourdon (early UNIX)
  • Martin Fowler, Ward Cunningham, Kent Beck (XP crowd)
  • Christopher Alexander, Eric Gamma, Gang of Four (Design Patterns)
  • 'Uncle Bob' Martin

machine layout, clusters, data flow, etc

Trainer: Hiram Clawson

loading a track, making an assembly

Trainer: Hiram Clawson

C libraries and GB code gotchas

Trainer: Jim Kent

kent src coding standards & libraries overview

Trainer: Angie Hinrichs

When we write library code or CGI code, that code will (hopefully) be in use for many years, and readability is extremely important because somebody else may need to debug or extend your code some day. Code is easier to read if it looks like it was written by one person as opposed to a jumble of different styles and indentation rules. Try to make your code look like Jim wrote it, and readers will thank you (or at least refrain from cursing you ;).

Basics of kent/src C formatting conventions:

  • multi-word variable or function names: camelCase, camel123Case (numbers and acronyms treated as words)
  • tab = 8 spaces
  • indentation = 4 spaces
  • opening brace { is indented on next line (not at end of line)
  • function declaration is followed by a brief comment describing inputs, outputs, gotchas like memory allocation details

If you use emacs, get Chuck Sugnet's jkent-c.emacs and add this to your ~/.emacs:

(load-file "~/jkent-c.emacs")

Higher-level conventions:

  • try to keep functions short enough to view in one editor screenful
  • empty lines between function declarations (but rarely within functions)
  • define functions before their first use in the file (to avoid duplicate declarations)
  • comment sparingly -- don't repeat what the code says, but say why you're doing something non-obvious
  • errAbort when there's an error condition -- much easier to find bugs that way

Get to know src/inc/common.h well (and common.c), and use its utilities and error-checking wrappers around C lib functions:

  • strcpy --> safecpy, sprintf --> safef, strncpy --> safencpy, strcat --> safecat
  • malloc --> needMem (and variants)
  • free --> freez, freeMem
  • memcpy --> cloneString, cloneStringZ, cloneMem
  • read --> mustRead, similar for write and close
  • verbose() for debugging comments
  • string utilities: sameString, startsWith*, stringIn, wildMatch, count*, chop*
  • sl* functions for list operations
  • sl* variants: slName, slRef, slPair

Other absolutely fundamental src/{inc,lib} modules include:

  • dyString: dynamically allocated, expandable strings
  • errAbort: context-specific handling of warnings and errors
  • hash: associate strings with any type of data
  • linefile: read in a file (or URL! and automatically decompress compressed files!) line by line
  • obscure: despite the name, this companion to common.h holds a lot of useful util functions
  • options: command-line option parsing

Below are several slide sets that provide a good introduction to Jim's programming philosophy and other local conventions:

  • Jim's Software Sermon (2002): PPT HTML
  • Jim's brief CGI programming intro (2007) - see esp. pages 6 & 7 about the cart PPT
  • Jim's Software Engineering & testing presentation (2012): PPTX
  • Jim's Locality & Modularity presentation (2012) PPTX

Others' overviews of kent/src:

  • Hiram's BME 230 presentation on GB, TB & kent/src (2011) - see esp. pages 5-20 PPT
  • Robert Baertsch's BME 230 presentation (2008): PPT
  • Angie's presentation to Bejerano Lab about working with and extending CGIs & utils (2008): PPT

GBiB, release process, development tools

Trainer: Max Haeussler

debugging tools

Trainers: Angie Hinrichs & Max Haeussler

Read the genomewiki page Debugging cgi-scripts

how not to break the build!

Trainer: Brian Raney


background reading material

hgFindSpec: http://genomewiki.cse.ucsc.edu/index.php/HgFindSpec

Our csh - bash equivalence document: $HOME/kent/src/hg/doc/bashVsCsh.txt

VI: http://genomewiki.ucsc.edu/genecats/index.php/VI_quick_start

Cluster Jobs: http://genomewiki.ucsc.edu/index.php/Cluster_Jobs

Retrieved from "https://genomewiki.ucsc.edu/genecats/index.php?title=Training_new_Browser_Engineers&oldid=4886"