• Unix bits
• vi quickies
• ACE and SIM operating stuff
• XML
• MARC and Z39.50
• Searching
• Whatnot  - basic tips, orientation, tricks
• Features - addons, Unicode, etc.

• Find a File Anywhere- down in a directory tree (below you), such as here with gunzip: find . -name 'gunzip' -print
• Get a line count of how many files/directories are in a directory-  ls -l | wc -l
• get a list of everything below me, and dump it in a file called "filename" - ls -lR > filename
• Run a job in the background from the beginning Such as with the Java string below, type all that in, then leave a space at the end, and an &
• After setting path, running XT: java (if memory, then add like -ms60M -mx400M depending on server) com.jclark.xsl.sax.Driver file.xml script.xsl out.xml
• ftp in local directory : After doing a run ftp domain.net, then lcd to c:\whatever to set local directory, run bin, and then do get/put.
• ftp multiple files : Set teh prompt to non-interractive by typing "prompt" and return, then use "mget *.*"
• Unix File compression
  1. tar cf archive.name.tar files (e.g., tar cf bigstuff.tar all*.*)
  2. Check what you have in its contents with tar tvf bigstuff.tar
  3. then to compress it, do "compress bigstuff.tar" and you'll get "bigstuff.tar.Z"
• recursively delete below you : rm -rf
• make a symbolic link to some other directory, called "simservers": ln -s /opt/sim/servers/ simservers
• concatenate some stuff into a file: cat file.txt file3.txt newfile.text
• How much space is free?  df -k
• Make a directory? mkdir name
• Find out what processes are running? ps -elf
• Find out how many resources one process is taking -elf | grep process
• Move or Rename a file? mv original new
• How much space does this file/directory take? - du -sk name
• Get rid of all stuff in folders below where you are
• Change permissions? chmod 777 sets all read, write, executable, and chmod -R 777 does to all in tree below you
• Find out the environment? env
• Change the current path for a session? alias thing cd /frequently/used/path
• Reload after a change to .cshrc? source .cshrc
• make and automated alias to another directory by just typing its name- change your .cshrc: alias what_I_type 'cd /directory/of/choice'
• protect yourself from "rm" mistakes add "alias rm "rm -i"" to your .cshrc; will prompt to ask if you are sure.
• set dos-like keys in Unix stty erase [then hit the "delete" key to set it, or bksp]

vi Quickies (vim too)

• move about esc, arrows --also--top/bottom of file esc, shift-g to end of file, esc, 1 shift-g to top of file
• delete what character the cursor is on- esc, x
• escape a character such as wehn replacing & with &, use this to do so -- \& to escape it -- use the \ after the / coming before the replace, such as: esc, : %s/&/\&/g
• delete to end of line use esc, .,$ d
• GO To line # just type the number, followed by a shift-g
• new line and start typing on it right away, from anywhere, without inserting hard return, start new below line you're on without splitting the one you're on - esc, o  then start typing
• copy a line - esc, shift-y,p
• cut a line - esc, dd
• remove ^M character esc, :1,$ s/control-v, ret//, then return. Control-v and return makes the ^M character.
• replace a character while in escape - esc, arrows to character, r, then whatever you want it to be
• search- esc, / then type the word sought
• search and replace in a big file - esc, : % s/find/replace/g  the s says to substitute, the g says repeat if twice on one line, a space comes after the :, and no spaces after the %, the s, or the /
• Replace with a hard-return This will find an xml tag, and insert a hard return before it (switch to "a" from "i" for putting it in afterward)
  • esc :map @ /^Mi^[#
  • -- the ^M comes from hitting control-v-rtn, and the ^[# comes from cntrl-v-esc --
  • what you're doing is making a macro, so you have to have it repeat, by next doing esc :map #@
  • test it first with esc :map # jk (tests it on just one line)
• save current file - esc, : w
• save current file with another name - esc, : w filename.txt
• save file with current name and EXIT - esc, shift zz
• open another file-  esc, : e
• open another file without saving current one esc, :edit!
• quit- esc : q

ACE and SIM Operating Stuff

• BASIC RULE:  initial capital letter in a script will be a class or type, initial lowercase will be a variable
• login as "sim": you will be able to become user sim by typing rlogin -l sim rosetta.emory.edu, once you In the home directory of user 'sim' (probably /opt/sim) create a file called '.rhosts' The contents of the file should consist of one line, the machine name followed by your login name. e.g. ========================== rosetta.emory.edu jrgardn ==========================
• To split or "do" your database: simace splitter.ace myfile.blu mysource.xml > myfile.load
• To put it "up"- simddl cms myfile.ddl
• where's the stylesheet? - /opt/sim/servers/advsearch3.0/formatter/thefile.ace
• How to remove a database: use a ddl script, have the line "DDL DROP DB "my database" - save as "drop.ddl, and the content manager in this case is cms - then at the command line, type $simddl cms drop.ddl.
• Alternately, write ACE that uses the ZAssociation dropDatabase function.  These methods do not remove dtd's and formatters (often these are shared in multible databases anyway.
• go under the hood: Go to the directory /opt/sim/servers/cms3.0

You will see a file called db-list. This file stores information about
the databases available in the cms (with the exception of IR-Explain-1
and IR-Extend-1).

Edit the file db-list.
Delete the two lines about Database_whatever.
Decrement the number of databases at the head of the file.
Save the file and restart the CMS.
 

• Starting SIM from command line-  cd into opt/sim/servers/boots3.0; then run simboots.run start; preferred method is to got to /etc/rc3.d and run S98SIM3.0 start as it makes sure all executables are owned by the right user.
• All standard SIM Ports (for access/admin/control in browser GUI):
• 7700 - directory server - port listings, protocols (http, Z39.50, etc.)
• 7702 - boot server (start and stop servers)
• 7704 - Security and logging server (set up who gets in to what/where)
• 7706 - Simple search interface
• 7707 - Advanced search interface
• 7708 - stylesheet wizard
• 7709 - database buildwizard
• abstract record structure is manipulated for multiple tag sets via DDL CREATE SCHEMA  Author  BIND "bib1-" "author" [newline] BIND "GILS" "origin"

 

• Buildwiz (7709 port) for XML-
1. copy /opt/sim/r3.0/dtd/xml8.soc into your buildwizard tmp directory and call it "my_xml_database.soc"
2. run the buildwizard as usual (assuming your dtd and source are in the buildwiz tmp directory, of course)
3. and then open your resulting .ddl file (called my_xml_database.ddl) in vi or emacs and get the line that says: "FragmentSGML" SGML MIME "text/sgml" PHYSICAL and make it read: "FragmentSGML" SGML OPTIONS "Catalog=my_xml_database.soc" MIME "text/sgml" HIGHLIGHTTEXT "<?SIM HI=\"ON\"?>" "<?SIM HI=\"OFF\"?>" PHYSICAL
• this just says, yes, it's sgml, but use this other catalog file, putting SP into XML mode
• alternately (untested), change it to: "FragmentSGML" XML MIME "text/xml" PHYSICAL
4. Then just run the usual simddl cms my_xml_database.ddl from the tmp directory

MARC & Z39.50

• Z39.50 - suports version 2 of Z39.50 well enough, and by default
•   version 3 defaults and the OID (Object Identifier) options for each Bib-1 set
• allows RPN QUery Type 1 (Rev. Polish Notation), and Query Type 2 (CCL, cf. ddl files)
• SimCLI learns via Explain Facility - you can use the line DDL CREATE CCLINFO for internal attention to CCL equivalencies or edits you are making
• abstract record structure is manipulated for multiple tag sets via DDL CREATE SCHEMA  Author  BIND "bib1-" "author" [newline] BIND "GILS" "origin"
• Full or Brief SUTRS returns, default for GRS-1
• Suppord for all Z39.50 facilities, even a well-fleshed Explain and Extend

Searching

• Fuzzy Searching - "title=@fuzzy(cut,90) -- this might give "cat"; where "90" is the accuracy in percents and "cut" is the parameter of the word sought
• stemming - "dog~" gives singular and plural (cf. Z39.50 truncation), all this is done with CCL/ISO 8777 syntax= Common Command Language)
• pattern matching - wom#n uses "#" to match any single character, and ? matches/allows more than onewildcard character
• Field searching - using CCL: TITLE=Smith (cf. "TITLE LIKE '%Smith%' " in SQL) -- but - also AUTHOR,SUBJECT=Smith, (not possible in SQL).

Whatnot

• Index Implementation- indexing only on indexed fields, records are not examined in queries, insuring performance - indexes are a B-tree and a postings file, the B-tree has multiple nodes,
• SIM indexes the vocabulary and #'s, this forms the B-tree, then maps where they are, forming teh postings; for scanning per field (cf. Z39.50 Scan facility)
• Databases can have multiple indexes with different levels and granularity, thus allowing concurrent markup and/or architectural forms-type solutions.
• It helps to have the line: /opt/sim/r3.0/bin $path in the .cshrc of the user/admin,
• Also in .cshrc add a setenv MANPATH \  /opt/sim/r3.0/man:$MANPATH
• a symbolic link to ace in the buildwiz helps, if put in opt/sim/servers for version 3.0.3: ln -s /opt/sim/r3.0/install/buildwiz/ace ace
• SERVER BITS and Parts:
• simwebs is the web user interface;
• simcms is the repository and retrieval (the core of it all, you can have multiple machines for your simcms),
• simdirs is the direcotry server telling other servers where everything is;
• simsls is the security and logging server
• simboots- bootserver starts first, and gets others going-- on UNIX only

SIM Goodies, add-ons, Unicode, features

• RDF Engine/webCrawler/Putator-  SIM does a web-crawl, looking in HTML for meta data, "title" tags, "h1" tags, etc.  This is pumped into a "putator" with a logic applicator (you can tailor this), to make the RDF tag set, via putative metadata (metadata generated implicatively), goes into a classifier (per topics, etc.) and then available through the cms.
• SIM Publisher- makes a standalone database instance on CD, fully runnable.
• SDI: Selective Dissemination of Information- included "out of the box" - notify of updates/changes in query history
• Unicode:
• All Java strings are 16-bit in string reading