Encyclopedia of Computer Science and Technology

446 sorting and searching• Discman CD-based music player (1984)• Handycam camcorder and Video format (1985)• Digital audio tape, or DAT (1987)• Blu-ray optical discSony would also become a major player in the consolegaming market (see gaming console). In 1994 the companyintroduced the PlayStation, followed by later modelsin 2000 and 2006. Sony is also a significant seller of digitalcameras, including the Mavica floppy disc (later CD), sincediscontinued. The company also introduced its proprietary“memory stick” for storage.Stumbles and SuccessesIn 2005 a controversy erupted when it was revealed thatSony music CDs included as part of their copy protection(see also digital rights management) a “rootkit” that couldallow PCs to be compromised. Sony eventually agreed withthe Federal Trade Commission (FTC) to exchange theaffected CDs and to reimburse damage to consumers’ computersthat might have occurred while attempting to removethe software. However, in 2007 a similar problem arose withthird-party software packaged with Sony memory sticks.Around the same time, Sony had to recall laptop batteriesthat had serious flaws that could cause them to overheatand catch fire. In 2006 Sony and Dell agreed to replace over4.1 million laptop batteries—this was followed by 1.8 millionSony batteries in Apple laptops and 526,000 in IBMand Lenovo laptops.Despite these setbacks, Sony continues to be very successful,with $70.3 billion in revenue and a net income of $1.07billion in 2007, and about 163,000 employees worldwide.Further ReadingLuh, Shu Shin. Business the Sony Way: Secrets of the World’s MostInnovative Electronics Giant. New York: Wiley, 2003.Nathan, John. Sony. New York: Houghton Mifflin, 1999.Sony America. Available online. URL: http://www.sony.com/.Accessed November 18, 2007.Sony Playstation. Available online. URL: http://www.us.playstation.com/. Accessed November 18, 2007.sorting and searchingBecause they are so fundamental to maintaining databases,the operations of sorting (putting data records in order) andsearching (finding a desired record) have received extensiveattention from computer scientists. A variety of differentand quite interesting sorting methods have been devised(see algorithm).Any application that involves keeping track of a significantnumber of data records will have to keep them sortedin some way. After all, if records are simply inserted as theyarrive without any attempt at order, the time it will take tofind a given record will, on the average, be the time it wouldtake to search through half the records in the database.While this might not matter for a few hundred records ona fast modern computer, it would be quite unacceptable fordatabases that might have millions of records.Sorting ConsiderationsWhile some sorting algorithms are better than others inalmost all cases, there are basic considerations for choosingan approach to sorting. The most obvious is how fastthe algorithm can sort the number of records the applicationis likely to encounter. However, it is also necessary toconsider whether the speed of the sort increases steadily(linearly) as the number of records increases, or it becomesproportionately worse. That is, if an algorithm can sort athousand records in two seconds, will it take 20 seconds for10,000 records, or perhaps five minutes?In most cases one assumes that the records to be sortedare in more or less random order, but what happens if therecords to be sorted are already partly sorted . . . or almostcompletely sorted? Some algorithms can take advantage ofthe partial sorting and complete the job far more quicklythan otherwise. Other algorithms may slow down drasticallyor even produce errors under those conditions.The range or variation in the key (the data field by whichrecords are being sorted) may also play a role. In some casesif the keys are close together, some algorithms may be ableto take advantage of that fact.Finally, the available computer resources must be considered.Today many desktop PCs have 1 GB (gigabyte) ormore of main memory (RAM), while servers or mainframesmay have several GBs. If the database is small enough that itcan be entirely kept in main memory, sorting is fast becauseany record can be accessed in the same amount of time atelectronic speeds. If, however, part of the database must bekept in secondary storage (such as hard drives), the sortingprogram will have to be designed so that it reads a numberof records from the hard drive in a single reading operation,in order to avoid the overhead of repeated disk operations.Most likely the individual batches will be read fromthe disk, sorted in memory, written back to disk, and thenmerged to sort the whole database.Sorting AlgorithmsThere are numerous sorting algorithms ranging from theeasy-to-understand to the commonly used to the exotic andquirky. Only the highlights can be covered here; see FurtherReading for sources for more detailed discussions.Selection SortThe simplest and least efficient kind of sort is called the selectionsort. Rather like a bridge player organizing a hand, theselection sort involves finding the record with the lowest keyand swapping it with the first record, then scanning backthrough for the next lowest key and swapping it with the secondrecord, and so on until all the records are sorted. Whilethis uses memory very efficiently (since the records are sortedin place), it is not only slow, but also gets worse fast. That is,the time taken to sort n records is proportional to n 2 .The selection approach suffers because on each pass thesort determines not only the record with the lowest key but