What are files and file systems

1
What are files and file systems

• Most files fall into one of three categories
• Executable files - applications and utilities
  like Netscape, Excel, Oracle,
• Executable files are further divided into
• applications - people use them directly for tasks
• system software OS, comms software, databases
• utilities conversion programs etc
• Data files - word documents, databases,
  spreadsheets, mpgs, jpgs, mp3, wav etc
• Support files - configuration initialisation,
  database indexes
• There are probably other types. These are not
  exact

2
Files - what and where are they

• Files are binary streams of data or instructions.
  
• They are stored on disc drives, CDROMs, DVDs,
  tapes and several other media
• From these storage media they are usually
  transferred, in part, to a computers memory
• They can be sent along telecommunication channels
  as streams of digital data through LANs, WANs
  and the Internet

3
Accessing files

• When we want to move a file from storage to
  memory, we open the file. The original file
  remains on the storage device.
• When we no longer want it in memory we close the
  file. This does not save changes that we may have
  made to the file.
• When making changes to data files, we must save
  the file before closing it. Saving updates the
  data in the original file in storage

4
File types

• Although all files are simply binary streams of
  data they behave quite differently and are seen
  as being different.
• When we open a wav or mp3 file we expect to
  hear sounds. When we open a jpg or gif file we
  expect to see a visual image. These
  representations are not intrinsic to the stream
  of data, they are only apparent because some
  application or other has been designed to
  represent them in that way

5

• When we open (or run) an executable file, the OS
  loads it into memory and begins to carry out the
  instructions in the file. (If it can - it may be
  the for the wrong OS)
• When we open a data file, the data is also loaded
  (partly) into memory, but the OS will present it
  to the user by first opening the application
  associated with that file.
• If a file contains text data but the OS does not
  know what application the file is associated
  with, the OS may open it with a text editor.
• Otherwise, the OS simply cannot open it .

6
OS, FAT and FMS

• Many operating system functions concern files
• OS use File Allocation Tables (FAT) to organise
  the storage of files on disc etc.
• If an application has a large number of related
  files, they may use a File Management System
  (FMS) to organise these. The FMS still uses the
  OS to save and open files etc.
• Databases are a special class of FMS

7
Naming conventions

• Some operating systems support file names in two
  parts - the real name and an extension
• In MSDOS based systems, files had names like
• excel.exe an exe file is an executable file
• letter.doc doc is a MSWord document (data)
• word.cfg cfg is a configuration file
• In some OS it is necessary to tell the OS which
  applications to associate with each file type so
  that the OS can open an appropriate application

8
Naming conventions 2

• A file extension is usually separated from the
  filename by a punctuation mark, often .
• In some OS, the names of directories are
  separated from the names of files by punctuation,
  often /
• Because of their special uses, these punctuation
  marks cannot be used in filenames in those OS
• e.g. Analysis/essay1.2 is not a valid filename

9

• Another way to tell the OS what type a file is,
  files often contain metadata i.e data which is
  not part of the real data but tells us about the
  real data.
• a Word document file contains the following
  metadata

_at_ñÿ N o r m a l mH lt A_at_òÿ lt
D e f a u l t P a r a g r a p h F o n t
ÿÿÿÿ
ÿÿ P e
t e r H y l a n d C \ H y l a n d D o c u
m e n t s \ t e s t . d o c ÿ_at_ t
_at_ _at_ G
T i m e s N e
w R o m a n 5
S y m b o l 3
A r i a l " ñ Ð h ÛcfÛcf
þÿ
àòùOh '³Ù0 x
Ì Ø ä ô
4 _at_ L X
h p ä test f
est Peter Hyland o ete
ete Normal y Peter Hyland o
1 te Microsoft Word 8.0 _at_ _at_
4oáltÕÁ_at_ 4oáltÕÁ
10
File Compatibility and standard formats

• Not all applications can read the metadata in a
  file
• If an application can read the metadata in a file
  it will still need to convert the data in the
  file into its own native format
• Some file formats, e.g. text files, are almost
  universally accepted by applications.
• Two special types of text files are
• space delimited - each word is separated by a
  space
• comma delimited words are separated by commas

11
More metadata

• Some files contain a lot of metadata or only
  metadata. These files are created in parallel
  with other data files about which they store
  metadata
• e.g. a database application might use 2 types of
  files
• a file containing real data about customers
• a file containing metadata about the data in the
  customer file.
• The metadata might include other names (aliases)
  by which the data in the customer file was
  described
• The number of characters in certain fields,
• access rights to some of the data fields

12
Buffering of files

• Often data files (and even some applications) are
  too large to fit into memory. When opened, only
  part of the file is transferred into memory.
• As the application needs more data it will open
  and close the data file, moving parts of it into
  memory to be processed, and then disposing of it.
  
• Similarly, communications devices and printers
  cannot load all of a file at once and so they
  store parts of it in buffer memory.
• Some operating systems use virtual memory (parts
  of the hard drive) to speed up file access

13
Storage mechanisms

• When whole files are stored on tapes or discs,
  they are either stored sequentially or for random
  access.
• Just like a video, tapes store their files
  sequentially - to get to a TV show at the end of
  the tape, you have to fast forward all the other
  TV shows first
• discs, like CDs are usually random access devices
  - you can jump to any track on a CD without
  playing the others first. You can jump to the
  start of a file
• This is about whole files and storage devices

14
File structures

• The data in a file may be a series of records all
  having the same structure but different details
• The data in the file can be arranged sequentially
  so that you must read all proceeding records to
  get to a specific record, lets say its record
  42.
• A sequential file can be stored on a tape or a
  disc. If it was stored on a disc, the OS could
  find the start of the file using random access
  but the database application would still have to
  search for records sequentially inside the file

15
File structures 2

• A random access file is structured in such a way
  that the application can locate any record and
  jump directly to that record, without reading
  preceding records.
• Random access files are most useful when stored
  on random access devices.
• If you were writing a database application, you
  need to know which file structure you are going
  to use and how to search it.

16
Directories / folders

• OS provide tools to organise files on storage
  devices.

Files are not stored in a physical location
corresponding to the directories. Files are just
linked to certain directories.
17
Deleting files and things in file

• When you delete a file, the file is not
  immediately gone. Usually, deleting a file
  tells the OS that the space occupied by the file
  is no longer in use.
• Many OS show this by putting the file in the
  Trash
• .
• The OS will use the space that was occupied by
  the file (write over it) whenever the OS needs
  to.
• Once the space occupied by a file has been
  written over, the file cannot be recovered i.e.
  once a file has been written over, you cannot
  take it out of the Trash

18
Fragmentation

• When a small file is deleted from a storage
  device, it leaves an available space - a hole if
  you like
• When the OS writes a new file to the device, the
  new file may not fit in the hole, so the OS puts
  part of the file in the hole and writes the rest
  of the new file in the next available space
• If the next space is not big enough, the file may
  be split across several spaces. The OS needs to
  keep track of where all the bits of the file are
  kept
• The new file is fragmented. Defragmenting a
  drive puts all file fragments together again -
  faster

19
Deleting records

• If a file contains database records, the
  applications that use the database may delete
  individual records from the file.
• Removing a record means a lot of work, so the
  application/database usually just marks (flags)
  the record as being deleted.
• When the application searches the file later it
  will ignore any records marked as being deleted.
• Periodically database files need to be purged to
  remove all the deleted records.

20
Backup copies

• Because deleting a file is eventually
  permanent, most OS make users confirm the delete
  command.
• Sometimes a storage drive will fail or crash
  and a necessary file will be destroyed.
• Make a backup copy of any necessary files, to
  protect from crashes and accidental deletion
• Keeping a sequence of backups is useful - if one
  fails you only lose the new data since the last
  backup e.g. essay1_1, essay1_2, essay1_3 etc.

21
Copying, moving and shortcuts

• When you open a file with an application and use
  the application to do a Save as, the
  application will create a copy of the file on
  another disc
• If you want a copy on the same disc you must
  either
• save the copy in a different directory OR
• save the file with a different name
• In a drag and drop interface, dragging a file
  to a new directory or disc usually moves the file
  i.e. you dont create a new copy, just move the
  original.
• If you want to be able to access a file quickly
  but dont want to move or copy it, use a shortcut