2
PRESS 'P' 10 QUIT THE ELUTION PROCESS Figure 2. Computer screen showing the gel tiltration column, lhe fractions being collected (left) and the relative 280-nm absorbance of lhe eluent with fraction number (right). XNOWNS : UWXNOWW : 11 Figure 3. Computer screen showing lhe plot of log (molecular weight) versus volume of effluent for knowns and unknowns. This should be the same graph that the stvdent generates The program is rich in simulated graphics (Fig. 2). The student can observe the Blue dextran and cobalt chloride as bands that separate as the chromatography continues. At the same time a window that shows a plot of eluent absor- bance at 280 nm versus tube number appears on the screen (Fig. 2). As in an actual laboratory fraction-collection pro- cess, the student can determine visually whether the main protein absorbing peak is eluted at or near the void volume or total volume. At the end of the elution process (or when thestudent ahorts the process), thestudeniisshowna menu that allows for the collection of the data or the alteration of experimental parameters. When the student determines that all appropriate data have been collected, the program writes the necessary information to a file along with the student's name for later review by the instructor. The information in each student file is gathered at the instructor's convenience by a separate program. The in- structor can call for graphs of the absorbance at 280 nm versus elution volume for both knowns and unknowns that are identical to those the student viewed. A graph of log molar mass versus elution volume (Fig. 3) (subjected to a least-squares analysis) can also be generated. This second program makes the simulation particularly easy to grade. Bond's Algorithm: Judging Correctness of Spelling Douglas Bond Riverside City College 4800 Magnolia Riverside, California 92506 Instructional computer programs require either numeri- cal or verbal response to questions. Although chemistry pro- grams usually emphasize the former, verbal response is re- quired in several of the programs, both commercial and private, that we have obtained over the years. Unfortunate- ly, we still receive programs that require absolutely correct spelling of the answer to avoid a negative response from the computer. We have observed that this can add to the diffi- cultv of the assienment since chemical terminolow is new to students, many-are not good spellers, and most are not good tvoists. Students tend to react unfavorablv to the comouter ahen an answer that is correct in all regids except fir the spelling is rejected. For many years, we have used a simple algorithm that checks for correctness of spelling in conjunction with several simple drill programs. I t was originally used with a symbol1 element-name drill program and has subsequently been used with others including programs where we are searching for keywords ina longer string. It is written in IBM PC Basic and can be adanted to otherswithout difficultv. Two string parameters are passed to the su1;routine-the correct answer (ARIGHT) and the student answer (AS- TUD). The subroutine then makes a direct comparison of the two strinm and if there is a match, the variable'lS'I'AT' is set to 1 (indicating a correct answer) and control is re- turned to the main program. If the two strings do not match, 'ISTAT' is set to 0 (appar- ently misspelled), and the subroutine makes as many as three passes through the two strings. In the first, the strings are compared character by character, and a note is made of the point of first difference, as well as total number of un- matched characters. If the number of errors is less than 25%, control is returned to the calling program. The assumption above is that the spelling error involves the substitution of one or more wrong characters for the correct ones. The algorithm then checks for one of three additional sources of error; (1) transposition of characters, (2) omission of a character, and (3) insertion of an additional character. If the number of incorrect characters in any of these passes is less than 25%. control is returned to the calling program with ISTAT still set at 0. If the number of incorrect responses is still greater than 25%, the answer is judged to be wrong and ISTAT is set to -1 and control passes to the calling program. We do not check for omission or insertion of two or more characters or any of the scores of other creative ways that students can misspell words. Although we probably miss some legitimate misspellings, it keeps the algorithm simple and the response time to a minimum. On an IBM PC with strings of 32 characters, the response time is virtually instan- taneous if there is a direct match, but may take up to 2.5 s if the computer must make all passes. Most input strings are much shorter than this, and students do not notice the slow- er response time. The algorithm is reasonably good at picking up most stu- dent misspellings. Part of the success of this is the per rent factor. which can easilv be chaneed (FRAC-line 9330). Set- ting i t much higher &an 0.30 60%) runs the risk of telling students that even garbage is simply a misspelled word. Setting it below 20% creates the probability of calling even a simple spelling error incorrect, especially if the word is less than eight characters long. Student response to the spelling check is nil. Whereas they complained if a correct .answer was called wrong simply 408 Journal of Chemical Education

Bond's algorithm: Judging correctness of spelling

  • Upload
    douglas

  • View
    213

  • Download
    1

Embed Size (px)

Citation preview

PRESS ' P ' 10 QUIT THE ELUTION PROCESS

Figure 2. Computer screen showing the gel tiltration column, lhe fractions being collected (left) and the relative 280-nm absorbance of lhe eluent with fraction number (right).

XNOWNS : UWXNOWW : 11

Figure 3. Computer screen showing lhe plot of log (molecular weight) versus volume of effluent for knowns and unknowns. This should be the same graph that the stvdent generates

The program is rich in simulated graphics (Fig. 2). The student can observe the Blue dextran and cobalt chloride as bands that separate as the chromatography continues. At the same time a window that shows a plot of eluent absor- bance a t 280 nm versus tube number appears on the screen (Fig. 2). As in an actual laboratory fraction-collection pro- cess, the student can determine visually whether the main protein absorbing peak is eluted a t or near the void volume or total volume. At the end of the elution process (or when thestudent ahorts the process), thestudeniisshowna menu that allows for the collection of the data or the alteration of experimental parameters. When the student determines that all appropriate data have been collected, the program writes the necessary information to a file along with the student's name for later review by the instructor.

The information in each student file is gathered a t the instructor's convenience by a separate program. The in- structor can call for graphs of the absorbance a t 280 nm versus elution volume for both knowns and unknowns that are identical to those the student viewed. A graph of log molar mass versus elution volume (Fig. 3) (subjected to a least-squares analysis) can also be generated. This second program makes the simulation particularly easy to grade.

Bond's Algorithm: Judging Correctness of Spelling

Douglas Bond Riverside City College

4800 Magnolia Riverside, California 92506

Instructional computer programs require either numeri- cal or verbal response to questions. Although chemistry pro- grams usually emphasize the former, verbal response is re- quired in several of the programs, both commercial and private, that we have obtained over the years. Unfortunate- ly, we still receive programs that require absolutely correct spelling of the answer to avoid a negative response from the computer. We have observed that this can add to the diffi- cultv of the assienment since chemical terminolow is new to students, many-are not good spellers, and most are not good tvoists. Students tend to react unfavorablv to the comouter ahen an answer that is correct in all r e g i d s except f i r the spelling is rejected.

For many years, we have used a simple algorithm that checks for correctness of spelling in conjunction with several simple drill programs. I t was originally used with a symbol1 element-name drill program and has subsequently been used with others including programs where we are searching for keywords ina longer string. It is written in IBM PC Basic and can be adanted to otherswithout difficultv.

Two string parameters are passed to the su1;routine-the correct answer (ARIGHT) and the student answer (AS- TUD). The subroutine then makes a direct comparison of the two strinm and if there is a match, the variable'lS'I'AT' is set to 1 (indicating a correct answer) and control is re- turned to the main program.

If the two strings do not match, 'ISTAT' is set to 0 (appar- ently misspelled), and the subroutine makes as many as three passes through the two strings. In the first, the strings are compared character by character, and a note is made of the point of first difference, as well as total number of un- matched characters. If the number of errors is less than 25%, control is returned to the calling program.

The assumption above is that the spelling error involves the substitution of one or more wrong characters for the correct ones. The algorithm then checks for one of three additional sources of error; (1) transposition of characters, (2) omission of a character, and (3) insertion of an additional character. If the number of incorrect characters in any of these passes is less than 25%. control is returned to the calling program with ISTAT still set a t 0. If the number of incorrect responses is still greater than 25%, the answer is judged to be wrong and ISTAT is set to -1 and control passes to the calling program.

We do not check for omission or insertion of two or more characters or any of the scores of other creative ways that students can misspell words. Although we probably miss some legitimate misspellings, it keeps the algorithm simple and the response time to a minimum. On an IBM PC with strings of 32 characters, the response time is virtually instan- taneous if there is a direct match, but may take up to 2.5 s if the computer must make all passes. Most input strings are much shorter than this, and students do not notice the slow- er response time.

The algorithm is reasonably good at picking up most stu- dent misspellings. Part of the success of this is the per rent factor. which can easilv be chaneed (FRAC-line 9330). Set- ting i t much higher &an 0.30 60%) runs the risk of telling students that even garbage is simply a misspelled word. Setting i t below 20% creates the probability of calling even a simple spelling error incorrect, especially if the word is less than eight characters long.

Student response to the spelling check is nil. Whereas they complained if a correct .answer was called wrong simply

408 Journal of Chemical Education

REH REIl """ .... "..'.'..*.... ..... """." .............. * ............ *.. n = IPN(ARIGHT1: FPAC-.25: ISTAT = 1: DEF PNSMREIPI - P/TL IF liSTUD = ARIGHT THEN RETURN

IF FNSWRE(H1SSBD) < -C THEN RETUlUl BMECK = &STUD amm = nia$jacmxrmsT.l) HID$lBCHECI(.FIRST,1) = HIDS(BCHECK,FIRST*I.U nro$(mrcn,rmsr*1,>) = LEFTIIA-P.~~ IF FIRST O TL THEN IF BCHECI i ARIGHT THEN RETURN

9376 TST&T = -1 9178 RETURN

Figure 4. Basic pmgram fw comparison of student responses wilh accepted answers including detection of simple misspellings.

because i t was misspelled, prompting the student that his answer is probably right, but the spelling needs checking, induces no gratitude. They seem to expect that the computer would detect spelling errors and given them the chance to make necessary corrections.

A listing of the subroutine is included (Fig. 4). The author expresses appreciation for helpful comments from the re- viewers.

Interfacing the TLC548 Analogue-to-Digital Converter

Bruce D. Westllng John Burroughs School

755 South Price Road St. Louis, MO 63124

Allan K. Seldel Lambelt-St. Louis International Airport Authority

Margaret E. Bahe John Burroughs School

Those who have been using the Apple I1 game port as an interface with theoutside world now have the opportunity to expand the range of physical parameters that may be stud- ied with the computer. Using the TLC548 analogue-to-digi- tal converter (ADC) i t is possible to measure the voltage output of specific ion probes, the differential voltage output, after amplification, of Wheatstone bridge circuits, and the voltage output of curreut-to-voltage amplifiers. Analogue data collection no longer needs to be restricted to game paddle input, which requires a resistance and associated current change from the transducer circuit.

Gome 110 Socket

Phone Jack / 0 0 1

Figure 5. Anaiogldigltai interlace circun tor use with Apple ii game port.

For the past few months we have been using the Texas Instrument TLC548 analogue-to-digital converter IC to col- lect laboratory data. This device is sold a t Radio Shack stores and costs $6.95. The literature provided with the TLC548describes it as. "a complete dataacquisition system on a single chip designed for serial interface with a micro- nroressor". The chin is built around an 8-bit switched-ca- r.......- ~ - - ~ ~ ~~~~ .~ ~ ~

pacitor successive-approximation ADC and includes its own system clock, which operates a t a 4 MHz.

Because of the Apple IIe game port design, which includes a Dower source, annunciator outputs, push-button inputs,

control inputs, and a strob;, i t is easy to interface the TLC548 with this computer. We combined the TLC548 with a 7404 inverter chip and a small software control program tu create an %bit voltage data sampler.

Firmre 5 shows the interface circuit. We use the wire wraD technique for joining circuit components together. ~ h f s method allows the builder to make auick connections. The TLC548 IC and inverter wire wrap sockets are glued to a small perfboard. After wire wrapping, this assembly is housed in a plastic case. Connection with the computer's game port is by way of a ribbon cable and 16-pin socket. Holes for thissocket are drilled in the case's lid; and then the

Table 2. Interface Parts List

Radio Shack No. or other source Cast

TLC548 analogtodigital converter 276-1796 $6.95 Hex inverter 7404 276-1802 $0.99 Project case 2% X 4% X 1% 270-222 $2.19 Perfboard microminiahre 2761394 $1.89 Wire wrapping socket: 8 pin 2761988 $1.39 Wire wrapping socket: 16 pin 2761994 $1.59 Wire wrapping socket: 14 pin 276-1993 $1.49 1bPin dip jumper cable 276-1976 $3.99 Wire-wrapping tool 276-1570 $5.95 30Gauge wire for wrapping 278-503 $2.39 Phone aids, modular jack Kmart $1.78

Volume 66 Number 5 May 1989 409