-
Add Bookmark | Recommend this book | Back to the book page | My bookshelf | Mobile Reading
Free Web Novel,Novel online - All in oicq.net -> Historic -> Lingao Qiming Chapter 325 Card Copy
Previous page Return to Catalog Next page
- So Fonnuo adjusted the data processing mode of the tabulating machine.
Tabulating machines were previously designed to read punched cards column-by-column. In this mode, even if it only takes 0.1 seconds to read each column, it will take 8 seconds to read 80 columns, which does not include the card transmission. time. After careful consideration, Fonnuo believed that in practical applications, not all columns of each punched card actually require statistical or cumulative operations.
Taking salary statistics as an example, it is impossible to have only salary-related data on a punched card. On the contrary, more columns should be used to store the worker's name, gender, department, job type, job number, etc. Information such as name location code, department code, job code, job number, etc. are not necessary for statistics.
Therefore, under the current conditions, it is enough to count only one piece of data at a time. Even if a piece of data is multi-digit and distributed on multiple columns of a punch card, it is feasible to only process one bit at a time. Anyway, The statistical results also need to be manually summed in the end.
In this way, when the tabulating machine processes each card, it only needs to read a certain column of numbers on the card and add them up, which is naturally much faster and saves more power.
Vonnuo spent another two days simplifying and modifying the device, and finally stabilized the tabulating machine's card processing capacity at more than 3,000 cards per hour, which should be able to meet the Senate's general statistical calculation needs. The only drawback is that if the data to be calculated has 3 digits, you need to go through the machine three times. It's still not convenient enough from an ease-of-use perspective. But now Fonnuo is trying to verify the technical ideas, and more improvements need to be made in the practical stage.
Having conquered the tabulating machine, in fact most of the mechanical mechanisms and relay control mechanisms have been successfully developed. Other types of machines are just changes in design, additions, subtractions, and combinations of components. In the punch card machine system, the sorting machine, which is second only in importance to the tabulating machine, is to change the relay used on the tabulating machine to control the dial and realize the accumulation function to separately control multiple card feeding mechanisms and control the Cards with holes for different numbers in the reference column are sent to different card pockets.
However, just as he was gearing up to take down the sorting machine in one fell swoop, a new problem arose.
Due to card quality issues, the card data that had been shipped were seriously damaged during the intensive testing of the tabulation machine in the past few days, and experimental materials will soon be in short supply.
Von Nuo had to put aside his eagerness to try the sorting machine and develop a copying machine first. With the copying machine, existing cards can be copied at will, and the problem of card shortage will no longer exist.
He found information about the IBM513 duplicating machine from IBM in the old time and space. Its full name is "Automatic Reproducing Punch". This is a model that only appeared in the 1940s. The principle is not difficult, but the structure of the machine is a bit complicated. Von Nuo does not intend to copy it exactly, but just build a prototype based on the principle.
He tried his best to draw a diagram on the paper as neatly as possible. The core component of the copying machine is a linkage transmission mechanism that can place double rows of cards. When the template card passes through the processing unit, the brush of the card reading mechanism sweeps the card, and at the punching position The connected circuit activates the relay in the corresponding position, thereby controlling the punching mechanism of the new card processing unit to punch holes in the same position of the new card. This time, reading and punching need to be done column by column.
However, Fonnuo got stuck when he was thinking about the control of the punching knife: there are only a few millimeters between the holes in each column on the punching card. He made the relay so big, how could he control multiple knife heads to punch holes at the same time? ? After thinking about it for a long time, I couldn't figure it out, so I had to go to the veteran of the mechanical port for advice the next day with a half-baked schematic diagram.
Still looking for Sun Li, as soon as he entered the office, he found that Zhan Wuya was also here. Feng Nuo had lived in the machinery factory for a long time and was familiar with Zhan Wuya. Knowing that he was a real person, he was not polite and polite. He directly spread the "drawings" and asked questions. Zhan Wuya laughed loudly, pointed to an English typewriter produced by Lingao next to him and said, "Go and see how that typewriter can type all the characters on more than 40 keys to the same position?"
Feng Nuo slapped his head, yes, he also had an English typewriter in his office, but he had never thought of this.
Zhan Wuya continued, "This is the most basic problem in mechanical design. In addition to using the arc type structure of a typewriter, there are many ways to solve it. Don't mess around with mechanical design by yourself. It¡¯s a matter of course. Let us handle this professionally. You should just design your relay control plan honestly. Yesterday, the Superintendent even asked about the progress of this project.¡±
Because Feng Nuo actually didn¡¯t understand mechanical design, the drawings he handed over to the machinery factory were basically just ¡°schematic diagrams.¡± He copied past design drawings for some ready-made structures, but when it came to the specific production process It still depends on the technical personnel of the machinery factory to redesign it.
Feng Nuo quickly reported on the progress and promised to speed up research and development in the near future. Please rest assured, leaders.
The copying machine did not pass the test as smoothly as the tabulating machine.
It was already evening when the prototype was delivered. First, we tested copying 1 and 10 cards respectively. After manual inspection, no problems were found. The copied new cards were exactly the same as the template cards. So Feng Nuo told Qian Yuzhi and Li Janai to try copying 1,000 cards at once.
Because it is copied line by line, the speed of the copying machine is roughly similar to that of the tabulating machine before optimization. It takes more than 3 hours to process 1,000 cards. Feng Nuo left work after explaining. The leader was very concerned about it, so Feng Nuo often worked overtime during this period. Today, he finally found a job that did not require any technical content, and he could just let Qian and Yuzhi keep an eye on it.
When copying a small amount of cards, you can check whether the perforations are consistent by overlapping the old and new cards and looking at the light. But 1,000 cards can no longer be checked for correctness manually. Therefore, Feng Nuo asked them to use a tabulating machine to check whether the perforations were consistent after copying. The method is to count whether the number of 0-9 in each column of the two stacks of cards is consistent. If the results are the same, the probability of a problem can be ignored. However, according to the current speed of the tabulating machine, the inspection time for each column is about 20 minutes, and 80 columns would take more than 20 hours, so two people had to work in shifts at night.
The next morning, Feng Nuo went to the Planning Institute for a small meeting and arrived at the office around noon.
He found that Feng Shan was leading Qian Yuzhi and Li Janai to check whether the punch holes of the cards were consistent one by one against the light. ¡ª¡ªAfter probably checking for a long time, the eyes of the three of them were a little blurry, and Qian Yuzhi and Li Janai even had dark circles under their eyes. Feng Nuo was very surprised and quickly went over to ask.
It turned out that Qian Yuzhi and Li Janai had been using a tabulating machine to check the copy results of 1,000 cards last night, but found that the old and new cards had different numbers of 0-9 in a certain column. The copied cards had one less 3 but more. A 4. In other words, the copying machine made an error and mistakenly typed 3 instead of 4 on a certain card. So they are now trying to figure out which card is at fault.
¡°The work attitude is commendable, but the working methods are stupid.¡± Feng Nuo commented. He walked over and picked up the two stacks of cards that the three of them had not checked yet, and asked, "How many have been checked?"
"122 pictures."
"114 pictures."
"107 photos."
The three people answered separately.
"Look, three people only checked more than 300 cards in the morning. In the future, we will test 10,000 cards. How many people will we need?" He handed a stack of cards to Feng Shan and said, "Check Come out with 330 cards."
Then he asked Qian Yuzhi which column was wrong and adjusted the position of the card reader brush of the tabulating machine. He also divided the remaining pile in his hand into two piles of 330 and 327.
Afterwards, Feng Nuo used a tabulating machine to count the 330 new cards that Feng Shan had found and the 330 old cards that he had sorted out.
Ten minutes later, the statistics ended, and the results were the same.
"So, the wrong card must be in the other half." Feng Nuo pointed to the remaining two piles of cards.
So he separated 160 cards from each of the remaining two stacks of 327 cards and entered them into the tabulation machine. This time the results showed that the wrong cards were among these 160.
Feng Nuo divided the batch of cards into two parts, each containing 80 new and old cards corresponding to one-to-one. He took a random part and repeated the above operation.
¡¡
A few minutes later, the suspect range was narrowed to 10. Feng Nuo distributed the cards to Feng Shan and three others, and quickly found the wrong card.
Feng Nuo put the wrong card on the table without looking at it. Instead, he said to Feng Shan: "The purpose of our making computers is to replace manpower in completing mechanical calculation work. It does the mechanical work without you doing it. What do you want to do? Think about how to make it work. Achieve your goal. Tell me, what is the principle of finding the wrong card just now?"
"Binary search method?" Feng Shan asked uncertainly.
Feng Nuo nodded, "It's different, but the idea is similar. If you think about it for a moment, why I asked Yuzhi and Kana to use a tabulating machine to check whether two stacks of 1,000 cards are consistent yesterday, you can think of many solutions. At least , you can enter 100 cards each time, up to 10 times, and you will also know which 100 cards the wrong card is in. - Of course, if you can think of this, you will also realize that binary search is the fastest. "
"The premise for us to do this is that the probability of copying errors is very low. Otherwise, if multiple punch cards have errors in the same column, this method will not work. So, yesterday we first checked the results of 1 and 10 copies. No problem."
"The facts also show that our estimate is consistent with the actual situation. 1,000 cards, that is, there are 80,000 column copy operations. The two of them checked more than 30 columns from last night to this morning. In fact, only one column was found to be wrong, and statistics found that there were only fewer There is a 3, and there is only one more 4. This is what you knew before looking for the error card, right? We have no other prior knowledge, so we can only assume that the error probability of the replicator is 1/30000, regardless of The current errors are caused by the accumulation of errors on multiple cards in the same column. Because the probability is lower, they can be ignored for the time being.
??¡ª¡ª
Next update: Volume 7 - Chapter 52 of the two-level guide??You knew about it before you looked for the wrong card, right? We have no other prior knowledge, so we can only assume that the error probability of the replicator is 1/30000. We do not consider that the current error is the accumulation of errors of multiple cards in the same column, because the probability is lower and can be ignored for the time being.
??¡ª¡ª
Next update: Volume 7 - Chapter 52 of the two-level guide - Didn't finish reading? Add this book to your favoritesI'm a member and bookmarked this chapterCopy the address of this book and recommend it to your friends for pointsChapter error? Click here to report
- Previous pageCatalogNext page