Power Programmierung

home *** CD-ROM | disk | FTP | other *** search

/ Power Programmierung / Power-Programmierung (Tewi)(1994).iso / magazine / c_news / 10 / cnews010.nws

Wrap

Text File | 1988-08-15 | 55KB | 1,139 lines

*------------------------------------------------------------------------* | | | C News - International C Newsletter, Compiler review, and | | Tutorial. | | | *------------------------------------------------------------------------* Volume 1, Number 10 9 Aug 1988 Editor | Editor's Corner: The Heap .................... 1 Barry Lynch | by Barry Lynch | Australian | Magazines: The Latest Crop ................... 2 Office: | by Barry Lynch David Nugent | | Book Reviews: | QuickC Programming for the IBM ................ 4 | Reviewed by Barry Lynch | | Database Design .............................. 6 | by Arnie Cherdak | | | Latest PD/Commercial Software Release ....... 19 | | Article Submission Standards ................ 20 | | Address's ................................. 21 | | Index ...................................... 22 | | Distribution Points ........................ 24 | | User Response Form .......................... 25 | | | | C News is an Electronic Journal published by the | C BBS in Burke, VA Bi-Monthly. All rights are | reserved, and all opinions expressed are of the | authors and any damage sustained, either physically | or financially via use of the routines in this | journal are at the sole expense of the reader. | | C News 1-10 Page 1 09 Aug 1988 ================================================================ The Heap: Messages from the editor ================================================================ COMPUTER AIDED SOFTWARE ENGINEERING: Some thoughts .... One area of programming that has always interested me, is computer aided software engineering. I suppose what I am really looking for is a software tool that helps me to automate the design, coding and maintenance of my code/applications. This need is not just for the PC world either, but the mini and mainframe world as well. According to Datamation < an DP industry trade journal >, the largest portion of a corporate MIS's group time and resources is spent on maintaining software that is in production. This I can agree on, from first hand experience. I work for a large corporation that has annual sales over the 1 Billion mark. I work in an Engineering group that is responsible designing, planning, projectizing, implementing, and maintaining application software on minis and mainframes. The biggest problem we face is: getting ALL of the functionality we request in a release, and getting releases on a timely basis. Both of the problems are directly related to the amount of work that is requested from the Corporate MIS groups, and the nature of the business we are in. Maintaining code is difficult, because of the myriad of relationships the various modules in an application have with other seemingly unrelated applications from other organizations. What is needed is the ability to link all applications together, so that changes in one that affect others, are reported on, and acted upon at the same time. What generally happens is that one department will make a change, that another affected department finds out about when the application starts to crash. Thus necessitating emergency coding by the affected group, thus delaying other projects in progress etc. I feel that the maintainability of code is so important that a message area here at the C BBS - Number 22 - has been dedicated to the subject. Other the coming months we hope to begin to understand this field more. If you are interested or involved in this field we would like to hear from you. HONEYMOON.. As some of you know I will be getting married on the 27th of August. Therefore, the C BBS will be on autopilot from the 27th to the 10th of September. So if you leave me a message in that time, do not get offended if I do not answer right away. I will not be taking a PC with me or have access to one. So enjoy.. C News 1-10 Page 2 09 Jul 1988 ================================================================ MAGAZINE Reviews: by Barry Lynch ================================================================ Magazine: Micro Cornucopia Issue: No. 43 Sept - Oct 88 Publisher: Micro Cornucopia Inc 155 NW Hawthorne, Bend OR 97701 Subject: The Grass-Roots side of the PC World. This issue of MC is dedicated to Database's. <Funny how this issue of C News is covering the same subject!> One of the things that I like about MC is it's approach - grass roots. MC reminds me of what Byte used to be in the old days - informative, and interesting. Quite frankly it is one of the few magazines that I look forward to getting. This issue contains: - Accessing Dbase III records from Turbo Pascal. - Building a C Database. - Selecting a DBASE III Compatible Compiler - Working with Paradox - Build a Pascal Database. - Designing Custom PC Cards. For those of you that read the FidoNet C_Echo on a regular basis, you will recognize the name of Scott Ladd. Scott writes the "C'ing Clearly" column for MC, and does an excellent job of keeping on top of the latest product releases, bug reports. This issue of MC, he digs into ANSI C some. Check out MC it is worth the yearly $18 fee for 6 issues. A tad more expensive than the other magazines that are available. But at least you get the feeling that the folks in Oregon like what they are doing, and want to hear from us users. C News 1-10 Page 3 09 Aug 1988 ================================================================ MAGAZINE Reviews: by Barry Lynch ================================================================ Magazine: Microsoft Systems Journal Issue: Vol 3. No 4. July 1988 Publisher: Microsoft Corporation 666 Third Avenue, New York, NY 10017 Subject: Articles associated with Microsoft Products. This issue of MSS includes the following articles: - DARWIN(tm) A Merrill Lynch Microsoft Windows Application - Codeview(tm) for Windows. - OS/2 Graphics Programming Interface. - MSC MASM 5.1 - Color mixing principles and how color works in the raster video model. - Creating user defined controls for your own Windows applications. - SQL server brings DBMS to OS/2... One of the problems with this publication is it's nasty habit of taking Microsoft press releases and making articles out of them. However, in among the chaff are a few grains of wheat worth digging for. If you are serious about Windows or OS/2 programming, then MSS is a good resource tool to have. C News 1-10 Page 4 09 Aug 1988 ================================================================ MAGAZINE Reviews: By Barry Lynch ================================================================ Magazine: Dr. Dobb's Journal of Software Tools Issue: #142 August 1988 Publisher: M&T Publishing Inc. 501 Galveston Drive Redwood City, CA 94063 Subject: Articles associated with Programming. First, the August issue of Dr. Dobb's is the annual C issue, and is a must for all C programmers. Second, Dr. Dobb's is still one of the best programming magazines around and occupies a permanent home of my book shelf. <Labeled Magazine holders and all..> This issue of DDJ includes the following articles: - C Compiler Review - Find that Function - Automatic Module Control in C - A tool for Secret Key Crytography - C Programming and others. This issue is a milestone of sorts, Allen Holub has been the C Programming columnist for as long as I can remember. He has stepped down from that position and been replaced by Al Stevens, of 'Programming TSR's ' fame. It will be interesting to see where Mr. Steven's takes the column. At any rate, DDJ is a magazine that all programmers should read on a regular basis. Some of the c source that has appeared in Dr. Dobb's has been published in a series of Books by M&T Press. These are excellent Intermediate-to-Advanced books and they are highly recommended. C News 1-10 Page 5 09 Aug 1988 ================================================================ BOOK REVIEW'S: By Barry Lynch ================================================================ Title: QuickC Programming for the IBM Author: Carl Townsend ISBN Number: 0-672-22622-7 Publisher: Howard W. Sams and Company 4300 West 62nd Street Indianapolis, Indiana 46268 The last thing I needed - and my checkbook - was to buy another computer book. For some reason If I do not buy at least one book a month, I feel like I am not keeping up with this crazy lunacy called the Computer Biz.. Not that I read all of the books I own, but you known what I mean. This is a book aimed at beginning users that use QuickC, hence the title. But, Mr. Townsend does something that I wish other authors would do as well. He provides a complete application in source that he uses to explain the C language. The application is a Mailing List program, that utilizes: Menus, a Database, and some reporting capabilities. I always felt that learning by example was one of the best ways to learn a subject. Mr. Townsend has done, a good job of just that. If you have just bought QuickC, and have thumbed through the manuals that Microsoft gave you. Go out and get a copy of this book, and sit down and relax. This is an excellent beginner's book. C News 1-10 Page 6 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ I never paid a lot of attention to the subject of DATABASE. It seemed that all I needed to know was how to run dBase II or III -- it didn't seem to matter -- in order to satisfy my fairly simple needs for putting some information together and then searching, sorting, and examining as required. It wasn't until I needed to do things with database files that were out of the ordinary that I became aware that there was more to databases than sequential files. A proper discussion of the subject of database could fill volumes and frequently has. In this article, I attempt merely to point out some of the subject matter of database technology and to focus on some of the more elemental aspects of programming for database using C. Lest you think I'm going to go off the deep end programming exotic and elaborate structures, be aware that I like to do things as simply as I can and that I'll use a ready made solution to the difficult part, a shareware package that takes care of the files and indexes and all that stuff that will allow me to focus more on the essence of the problem. This has an added advantage in that it will make this article easier to write and easier to read with less obscure detail. Let's get on with it. Starting from the top...a database is a collection of information, somehow related. A Database Management System is a collection of programs that is used to search, sort, and examine that database. Simple so far...but let's go on. Much of the software technology centered around databases is associated with ways in which information can be extracted from the data. Structured Query Language or SQL is the latest and greatest approach to programming searches of a large class of databases. Those of you who have experienced dBase may remember the LOCATE command that serves a similar purpose. Many database manager software packages offer a selection of subroutines (for the FORTRANners among us) that can be linked with users' programs to search the database for the information of interest to the user. More about searching and ways to do it, later. A closely related subject to searching for information is displaying the information resulting from the search. Although it's considered by some to be a more mundane aspect of database technology, the report probably commands more attention by the database user and by the programmer than most other aspects. The user interface, too, gets a lot of attention from the programmer as it does in most other types of software. C News 1-10 Page 7 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ These two aspects of database software are subjects that are too big and important to be discussed briefly so I won't even begin. I'll leave them for a future article. At a lower level of detail, frequently hidden from the user, is the fundamental structure of the database which includes the organization of the data and the basic mechanisms used to enter data into the database and to access that data. Since there is generally more data in a database than can be held in primary memory at one time and the data is generally required to be non- volatile, the database consists of one or more files stored on secondary memory (disk). Databases can be distributed far and wide and accessed by way of complex data communications means. However, this approach is not all that common and a database located at a single location is the rule. Certainly, this is the case for the small, PC-based databases we consider here. Database files can have a wide range of forms. At one end of the spectrum are complex files containing several segments that make up the entire database including directories, headers, indexes, and the data. These complex files can even be organized as trees with pointers to adjacent records (i.e., tree nodes) in the tree structure contained in each record. At the other end are simple, sequential or "flat" files with uniform records. DBase III, for example, is somewhere in the middle of the complexity spectrum with a two-segment data file containing a description of the data records at the front followed by uniform data records and a separate file containing an index. Depending upon the nature of the database management software, there can be one or more files that make up the database. The simplest case has the file composed of fixed size records, each record composed of fixed size fields with the field being the smallest element of data in the database. Indeed, if we consider the database, itself as a file for our present purposes, the art of database programming is really the art of programming searches for the information contained in the data file. This is the subject of the rest of this article. DATABASE 101 -- Linear Searching Let's continue with the simplest case. Suppose we had a list (file) of names and addresses for a small community of about 200 families. Furthermore, suppose we wanted to have a list of families living on Main Street. A very simple way to obtain that list is to visually examine the list of all the families from beginning to end, C News 1-10 Page 8 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ one family (record) at a time and to write down the family data for each family satisfying the criterion that its street name was Main Street. This is not a wonderfully sophisticated way to get the information we want but it certainly is effective. I just described a search method that could be programmed in C or in some other programming language to search through a file containing the family records to create a printed report. The process I went through is as follows: 1) Open the file to the first record. 2) Look at the street name for the record in view. 3) If the street name is Main Street, copy the record contents to a list. 4) If there are no more records, close the file and quit. 5) Otherwise, get the next record. 6) Go to step 2 and repeat the process. For those of you who haven't written such a linear search program before, take the time to write a computer program to execute these 6 steps. Use the data file, FAMILIES.DAT, included with this issue of CNEWS. You could use the printer to display your list or have your program create another file containing the Main Street Report. The record structure in the data file is as follows: struct { char name[25]; char street[25]; char street_number[10]; char other_data[68]; } family; Note that there are 191 records in the file and there are 12 families that reside on Main Street. Oh, yes, I almost forgot. The City Council was very budget conscious last summer when this file was created and they employed a high school student who was given the job to create this file. The student hadn't taken the first course in typing yet and, as a result, the fields in the file aren't as neat as they could have been. You may want to spend a little time writing defensive code to be able to read the fields reliably. Sorry about that but the database business IS full of things like messy data. <GRIN> C News 1-10 Page 9 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ DATABASE 201 -- Indexes Well, that's pretty good if all you want to do is generate a Main Street List from a 200 record file. What if you wanted to do some more complex things, say, create a sorted listing or work with a file having more records, perhaps several thousand records? What if the records are significantly larger? As the database gets bigger, it takes longer and longer to do things with it and if you're like me you don't have the fastest cpu and hard disk in town. There has to be a better way, a faster way to get what you want from the database than to read it record-by-record and search each record for the strings or numbers you want. There is a better way, it's called indexing. The idea of indexing is fairly straightforward. Create a file containing index strings or keys that are created from the database record fields that you would use during a linear search to decide whether you had reached a desired record. For the Main Street Search, you might want to use the street name field as the key. The index file record, then would consist of two fields, the key and the record number of the database record that the key came from. The index file would then be much smaller than the database file; having the same number but much smaller records. It would then take much less time to manipulate the index file. When the index is searched and a desired key is found, the database record can then be accessed immediately using the record number from the index multiplied by the record size to get the offset into the database file where the desired record begins. DOS file positioning commands can then be used to go directly to the desired record. Alternatively, you can use the C function, ftell(), to obtain a pointer to the start of each record before it's read. Please understand, if you have a specific item you're looking for and you only want to do it one time, you may want to do just the linear search through the database file because, to create an index, you need to read through the file at least one time anyway. The index to a database is useful if you are going to access the database numerous times and search for more or less randomly specified values of the index. There are other advantages to using the index file. It can be sorted to reduce the search time still more. If you sort alphabetically and search for Main Street, the search is over when the street names beginning with 'M' have been exhausted. Adding full records to the database can be done chronologically at the end of the database file. When the new index file records are added to the index file, it takes a lot less time to re-sort the index file than C News 1-10 Page 10 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ the database file. Similarly, more than one index file can be created to serve different purposes while keeping the chronological sequence of the database intact. I haven't said anything about the organization of the index file. It can be a flat file in the simplest case or, to make the searches faster, it can be organized in other ways with tree organization probably the most often used. Index organization is discussed below. DATABASE 301 -- More Sophisticated Ways There are ways to reduce the amount of index manipulation required by linear searching. Some of these include Indexed Sequential searching, tree searching, binary searching, and hashing to name just a few. Indexed Sequential searching is a technique in which the database is sorted and the index contains only a fraction (about 10 percent) of the number of records that exist in the database. Finding a record is done by locating a key close to the desired key in the index and performing a linear search in the database through just a small number of records near the key record. This is very effective where the number of database records is very large and manipulation of a regular index file would be too time consuming or the full index won't fit in available memory. Hashing is an approach that attempts to minimize the number of comparisons made of the desired key with the keys in the index in performing the search. A single access is the goal of hashing. With this technique, a Hash Function converts the desired key to an integer value which is interpreted as a table or array index or as an address. The value at that index or address is then a pointer to the desired data record or even the record itself. More on Indexed Sequential searching and Hashing in a future article. Binary searching permits rapidly arriving at the index record when using an index that is sorted in order of the key value. (NOTE that even string keys have a "value" when compared with other string keys. See the string comparison function descriptions in your programming language manual). Finding the desired database record is done by comparing the desired key with the key at the center of the index file. If there's a match, go right to the database record. If the desired key is larger than the center index key, look at the index key half way between the current index file location and the end of the index C News 1-10 Page 11 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ file. If the desired key is smaller, look at the index key half way toward the beginning of the index file. Continue this process until the desired index key is found or until you have gone through the entire file unsuccessfully in which case the search has failed. The binary search function looks like Listing 1. ------------------------------------------------- LISTING 1. BINARY SEARCH FUNCTION char *k(x); /* returns pointer to key for record x */ void binary_search(long n, char *key) { /* n = number of records in list */ long low,mid,hi,x; /* key is desired record key */ int found; found = FALSE; low = 1; /* beginning of list */ hi = n; /* end of list */ while(low<=hi && !found) { mid = (low + hi) / 2; /* calculate middle of list */ if(strcmp(key,k(mid)) == 0) /* k(x) looks up key value */ found = TRUE; /* match is found */ else /* no match, larger or smaller key */ if(strcmp(key,k(mid))<0) hi = mid - 1; /* key is < middle key */ else low = mid + 1; /* key is > middle key */ } /* end while */ if(found) /* return record number found */ return(mid); /* success -- record found */ else return(0); /* search failed -- record not found */ } /* end binary_search */ NOTE: 1.If the index list is in a file, then function k(x) will have to access that file to get the desired key values. 2.It is assumed that the index list is sorted in ascending order with index record n greater than index record 1. ------------------------------------------------- C News 1-10 Page 12 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ The binary search is just about the most efficient way to search a linear, sorted list whether that list is in memory as implied by LISTING 1, or in a file. After all, each time a decision is made to search in one half of the places (records) available, the other half is excluded and doesn't have to be searched. You can narrow the options quite rapidly with this technique. Note that if the database is in chronological order and the date-time the record was added to the database is included in each record, then a binary search of the file based on date-time can be performed without the use of an index. EXERCISE: Write a function to create an index list of the FAMILIES file in memory. Use a structure for the index list element as follows: struct { int recaddr; /* FAMILIES record pointer */ char key[25]; /* street name */ } index_rec; Adapt the binary search function shown in LISTING 1 to your index. Write a k(x) function to return the key given a list position. Write a program to print the Main Street List using the above functions, and the index list in memory, and any other functions you might need to get all twelve Main Street records in the file. C News 1-10 Page 13 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ --------------------------------------------- DATABASE 401 -- TREES OKAY! You're on your way to becoming an expert or at least more familiar with the subject of programming database search routines. Next, with trees. What's a tree? You ask... Well, a tree offers a way to organize data to make it easier to search for particular data elements. A tree is a data structure constructed of nodes and linkages among nodes (the notion of the linkage as distinct from the node isn't really right but I'll correct that in just a bit). A tree is an orderly arrangement of nodes interconnected by linkages. In C parlance, the node is a structure containing a value or a set of values and two or more pointers to structures of the same type as the node. struct node { char key[25]; int recno; struct node *lptr; struct node *rptr; }; The structure depicted above may be created using the calloc() function as described in K&R. Note that the linkages among nodes are not separate entities but are in truth the pointers contained within the node structure. In the above case, lptr and rptr are the linkages between a node and its adjacent nodes in the tree. A tree is created at its root and it grows from there. In this case, a root is a pointer to a node. Define the pointer thus: struct node *root; Now, create the first node using calloc() which returns a pointer value. Assign that pointer value to root using a cast to get the pointer type correct. root = (struct node *) calloc(1, sizeof(struct node)); Note that root->lptr and root->rptr are of the same type as the variable, root, and that they may be used with calloc() in the same way to point or link from one node to another in the tree. C News 1-10 Page 14 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ root->lptr = (struct node *) calloc(1, sizeof(struct node)); In this way, new nodes are generated and linked to previously generated nodes, hence to the root of the tree. If one wanted to reach any node in the tree, it is a relatively straightforward matter to start at the root and follow the pointers until the desired node is reached. Welllll...it's a bit more difficult than that but not a lot more. However, it does get somewhat complex and, as I promised earlier, I'm going to keep this as simple as I can for you and for me. A tree has a visual appearance similar to Figure 1. ------------------------------------------------- FIGURE 1. Basic Tree Structure |pointer to root node | Root Node / \ / \ Node1 Node2 / \ / \ / \ / \ Node3 Node4 Node5 Node6 / \ / \ / \ / \ etc................ ------------------------------------------------- The tree of FIGURE 1 is called a Binary Tree because each node is linked to two other nodes further down in the tree. If each node contains a key and a corresponding record's file address, then the tree can represent the entire index to the database file. Furthermore, if the keys and addresses are installed in the nodes with the rule that keys having values lower than the value of a key in a node can only be installed in a new node pointed to by a left-directed pointer and higher valued keys go into nodes pointed to by right-directed pointers, then the tree can be searched for a key value with similar results to a binary search on a linear, sorted list or file. WHEW! That was a mouthful. I don't offer proof of the above but if you assume that the root is very close to having the middle or median value key in the database, then it becomes apparent that if you search for a C News 1-10 Page 15 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ specific key value starting at the root, and, for a key value lower than the root key value, you follow the left-directed pointer, you will never have to follow any pointers linked to the right-directed pointer. This is directly analogous to the binary search described above. Well, that's all well and good but what does it do for me and you? The value is not in the useful but simple binary tree. It is in what are called Multiway trees or trees that have more than one key and more than two pointers per node. These are also called B_Trees. These trees are searched in much the same manner the binary tree is searched with the notable exception that at each node containing more than two pointers to lower level nodes in the tree, there are more than two paths that can be followed and an appropriate number of decisions to be made to select the right path. A special case of a B_Tree, a B+_Tree, has special properties which, in addition to the capability to be searched rapidly for a random selection of key values, also lends itself to linear searching. Well, this is getting close to my maximum confusion index so I'll simply state that there is reason for wanting to use structures having this kind of complexity and that reason is search performance compared with some of the other search techniques mentioned above. Needless to say, programming such structures and the programs needed to maintain and search them is not trivial. It is probably wise to decide whether it is more important to solve the database problem for yourself or to pursue the education needed to create such programs. For me, the choice is obvious. I'll take the easy way and let someone else do it for me. If you're just dying to learn all you can about trees then try some of the books listed in the bibliography at the end of this article. For present purposes, I refer you to the .ARC file, BPLUS11.ARC, included with this article. This is a shareware package that does all this tree stuff for you and makes using trees for database applications a whole lot easier. This isn't the only such package. There are lots of records management support packages available in shareware and as regular products for sale that use the tree as well as other methods. The following is excerpted from the B-PLUS documentation contained in the ARC file. It would be appropriate to read it all before you use the package. The documentation file also includes a description of the parameters for each function as well as its source code. Note that I have version 1.1 which isn't quite right C News 1-10 Page 16 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ for Turbo-C, version 1.5. I spoke with Larry Hunter whose company markets this product, and he told me that BPLUS version 1.1A solves that and another problem. It's available on COMPUSERVE, now. In the meantime, Larry told me how to fix the Turbo-C 1.5 compatibility problem which I did. I've included a brief note on this along with the BPLUS11.ARC file in file, BPLUS11.NOT. "The B-PLUS File Indexing Module contains twelve functions that handle the retrieval of data records by key value. The keys that are used to locate records are null terminated strings. The data structures and constants that are used are defined in the header file bplus.h." The following are the names and descriptions of the twelve functions: open_index: The open_index function is used to open and initialize an existing index file specified by name and prepares the file for subsequent reading or writing. make_index: The make_index function is used to create and initialize a new index file specified by name and to prepare the file for subsequent reading or writing. close_index: The close_index file function clears the internal cache buffer and closes the specified index file. find_key: The find_key function searches the index file for the key value. locate_key: The locate key function searches the index file for the first key value which is equal to or greater than a specified key. add_key: The add_key function adds new entries to the index file. delete_key: The delete_key function deletes entries in the index file. first_key: The first_key function positions the index file pointer to the beginning of the index file. The function next_key can then be used to list the file in key order. C News 1-10 Page 17 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ last_key: The last_key function positions the index file pointer at the end of the index file. The function previous_key can then be used to list the file in reverse key order. next_key: The next_key function returns the next entry in the index file. Before next_key is called, the index file pointer should be positioned by the function first_key, find_key, or locate_key. Next_key can be used to locate the desired data record when duplicate keys are allowed. Another use is to process files, sequentially. prev_key: The prev_key function returns the previous entry in the index file. Before prev_key is called, the index file pointer should be positioned by the function last_key, find_key, or locate_key. Prev_key can be used to process files sequentially in reverse order. find_exact: The find_exact function searches the index file for a specific key and data record position. After playing with BPLUS for a while, I find that the only real criticism I have is that there isn't a tree-search function for multiple identical keys. The example program called DTEST.C that I've provided, prints out the Main Street Report, but it searches linearly through the index file to find the 12 Main Street records. There are other ways to use the BPLUS functions but this is the way I found first to make the example work. Just the same, BPLUS works well. ------------------------------------------------- BIBLIOGRAPHY Borland International, Inc., Turbo C documentation, Version 1.5, 1988. Borland International, Inc., "Turbo Pascal Database Toolbox". Hunt, "The C Toolbox", Addison-Wesley, 1985. Hursch and Hursch, "SQL The Structured Query Language", TAB Books, 1988. Kernighan and Ritchie, "The C Programming Language", Prentice- Hall, 1978. C News 1-10 Page 18 09 Aug 1988 ================================================================ DATABASE IN C: By Arnold Cherdak ================================================================ Schwaderer, "C Wizard's Programming Reference", Wiley Press, 1985. SoftCraft, Inc., "Btrieve User's Guide" Tennenbaum and Augenstein, "Data Structures Using Pascal", Second Ed., Prentice-Hall, 1986. Tsichritzis and Lochovsky, "Data Base Management Systems", Academic Press, 1977. Wirth, "Algorithms + Data Structures = Programs", Prentice Hall, 1976. C News 1-10 Page 19 09 Aug 1988 ================================================================ Latest Versions of Commercial/PD Software. ================================================================ Compilers: Ver: --------- ---- Microsoft C 5.10 QuickC 1.01 TurboC 1.5 /* With a rumored 2.0 out */ Datalight 2.1 MIX C ? PowerC ? Lattice C ? Watcom 6.0 C News 1-10 Page 20 09 Aug 1988 ================================================================ ARTICLE SUBMISSION STANDARDS AND ADDRESSES ================================================================ As I have repeatedly stated in this newsletter and previous issues, I would like to see user-submitted articles, reviews or questions. Listed below are the standards that should be followed to make my job easier as an editor. - Articles should be submitted in a ASCII non-formatted file. (Margins 0-65 PLEASE) - If the article include code fragments as examples. Then you can include the entire source file if you like for inclusion with the newsletter. - Book or magazine reviews should follow the same format, that is outlined in this issue. The publisher, author, title, and ISBN number are a must. - Compiler/and or product reviews, should include the version number and manufacture. If possible, reviews should include a sample program with benchmarks. If you have any questions you can contact me at the address's included on the next page. C News 1-10 Page 21 09 Aug 1988 ================================================================ ADDRESSES ================================================================ The C BBS is located at: C BBS % BCL Limited P.O. Box 9162 McLean VA, 22102 or you can send netmail to: 1:109/713 C News 1-10 Page 22 15 Jul 1988 ================================================================ INDEX ================================================================ Subject: Issue: Articles: Additional Comments of Filename Wild.. 6 Beginning C Functions 7 C Spot Run: A User Supported Library 7 Database Design in C 10 Filename Wildcard Expansion in MSC 4 Integrated Environment: TC & QC 5 Programming the Hercules Graphics Card 8 Programming the Hercules Graphics Card: Part 2 9 Talking with a Fossil 5 TurboC and Interrupts: A few Questions 2 Book Reviews: C Chest: and other treasures. 6 C Database Development 1 C Programming Guide 1 C Programming Language 1 C Programmer's Guide to Serial Communications 3 C Programmer's Library 1 C Primer Plus 1 C the Complete Reference 2 Crafting C Tools for the IBM PC 2 Learning to Program in C 1 Microsoft C Programming on the IBM PC 1 MS-DOS Developer's Guide 4 Programming in Windows 3 QuickC Programming for the IBM 10 Reliable Data Structures in C 1 TurboC: Memory Resident Utilities 5 TurboC Programmer's Reference Book 2 Compilers: QuickC 1 Software Reviews: Bplus11.arc 3 C_Dates.arc 4 Cdate.arc 4 Casm.arc 3 C-subr.arc 4 C News 1-10 Page 23 09 Aug 1988 ================================================================ INDEX ================================================================ Docu.arc 3 Ezwind.arc 8 Jcl-src.arc 4 Mscpopup.arc 3 Ndmake41.arc 4 Nuc-subr.arc 3 Prndoc.arc 6 Sed.arc 6 Shift_c.arc 4 Sysact11.arc 4 Tp_to_qc.arc 3 Xenixarc.arc 4 C News 1-10 Page 24 09 Aug 1988 ================================================================ DISTRIBUTION POINTS ================================================================ Board Name Number Net/Node Sysop United States C BBS (703) 644-6478 1:109/713 Barry Lynch Burke, VA Jaz C-Scape (904) 724-1377 1:112/1027 Tom Evans Jacksonville, FL Eastern C Board (201) 247-6748 1:107/335 Todd Lehr OTHER BOARDS THAT CARRY C NEWS: Exec-PC (414) 964-5160 .. Bob Mahoney Milwaukee, WI TAMIAMI (813) 793-2392 Gerhard Barth Naples, FL Sound of Music (516) 536-8723(2400) Paul Waldinger (516) 536-6819(9600 Hayes V) Canada Another BBS System (416) 465-7752 1:148/208 Mark Bowman Toronto, Canada Europe Fido_N1_1 31-8350-37156 2:500/1 Henk Wevers The Netherlands Australia Alpha-Centuri BBS 011-61-3-874-3559 3:632/348 David Nugent C News 1-10 Page 25 09 Aug 1988 ================================================================ USER RESPONSE FORM ================================================================ This form will be included as a regular feature in all future issues of C NEWS. What did you think of the content of this Issue? _____________ _______________________________________________________________ What improvements can you think of that would make C News a better tool for the C Community? _______________________________________________________________ _______________________________________________________________ What is your favorite section or sections? ___________________ _______________________________________________________________ What don't you like about C News? ____________________________ _______________________________________________________________ Additional Comments: _________________________________________ _______________________________________________________________ _______________________________________________________________ _______________________________________________________________