Author: Dann Corbit
Date: 18:30:08 09/02/05
Go up one level in this thread
On September 02, 2005 at 21:03:08, Uri Blass wrote: >On September 02, 2005 at 19:35:21, Dann Corbit wrote: > >>On September 02, 2005 at 13:29:11, Uri Blass wrote: >> >>>On September 01, 2005 at 22:00:52, David Mitchell wrote: >>> >>>>On September 01, 2005 at 19:57:26, Uri Blass wrote: >>>> >>>>>On September 01, 2005 at 16:50:38, David Mitchell wrote: >>>>> >>>>>Note that I am one of the members of the Israeli >>>>>chess rating committee(the committe is not only about rating but rating is one >>>>>of the subjects that they deal with it). >>>>>Members are all volunteers and we do not get money for it but it is important >>>>>for me to have a better rating system and it is clear that without a new >program it will not be easy to make changes in the rating system. >>>>> >>>>>I also have part of the data because GM Ram Sofer(the person who is resposible >>>>>for using the software that they have in the last months) sent me some access >>>>>files. >>>>> >>>>>Personally I know almost nothing about access so I simply translated the >columns to text files by copy and paste and wrote some program in C to >calculate statistics about the text files. >>>>> >>>>>I think that it is even better if the Israeli chess organization get rid of all >>>>>the access and simply save the results in text files and use some C program to >>>>>read text files and analyze them. >>>>> >>>> >>>>Trust me, Uri. You don't want to do that. You need a spreadsheet or a database >>>>program, or both. Too many benefits to list them all, but having different >>>>people work in diverse locations, with "some C program", is not what you want. >>>> >>>>Use Excel for a day, and you'll never want to go back to anything like this. You >>>>should be able to pick up an older (but perfectly good), version of Excel, by >>>>some company that has decided to upgrade to the latest version. >>>> >>>>Should be able to get it for very little money. >>>> >>>> >>>> >>>> >>>> >>>>>In that case if there is a mistake people only need to edit the text files. >>>> >>>>Oh yeah, there will be mistakes, count on it! >>>> >>>>> >>>>> >>>>>Note that I already have a program to analyze text files but the format of my >>>>>text file today is not good for editing because of the way that I copied the >>>>>data. >>>> >>>>There you go! No "homegrown" program will have half the features you will really >>>>want, or be half as easy to train others to use. >>>>> >>>>>I copied the data from the access file by copy and paste so I copied long >>>>>columns that is clearly easier than copying small lines that have only details >>>>>about a single game. >>>>> >>>> >>>>You should just use the Export data function. You can export your data in >>>>several formats, including ascii (text). >>> >>>Thanks for your advice >>> >>>Note that only now I am near a computer that can read the access files because >>>access is not installed in other computers but using the export data function is >>>very good idea. >>> >>>It is a different format but I get all the table as a text in few seconds >>>instead of getting it in few hours. >>> >>>If I understand correctly you say that using Excel is better than using access. >>>I did not know that it is so easy to translate access files to excel files. >> >>Export from Access. It will take you two minutes. > >I already found it and translated the data both to text files and to excel files >and sent the data to myself by gmail so I can read it from every computer. > >> >>I do not think Excel will be easier than Access. >> >>Leaving the data in MS Access will be better because you can apply rules to a >>database and you cannot apply rules to a spreadsheet. >> >>I have seen multi-million dollar disasters caused by using spreadsheets where a >>database should have been used. > >What was the problem? The project was called "Budget Workbench". The idea was that people all over the world would fill up spreadsheets with data and then the spreadsheets would be used to put data into a database. Of course, once I understood what they were trying to do, I knew that the project was hopelessly doomed. You cannot have thousands and thousands of freeform spreadsheets collected and expect that the integrity of the data contained in the spreadsheets is perfect. It is obviously going to fail. Now, with a database, you will have the same kinds of problems but they are self correcting. Suppose, for instance, that I have a database of chess players. One of the chess players is named Uri Blass. I want to enter some new game results for Uri Blass and so I try to add a record but spell the name incorrectly as 'Uri Balss.' The data entry form can have a trigger on it or a relation on it or a rule on it to recognize the problem and throw an error. Well, if there are 10,000 typing mistakes, we still get 10,000 errors thrown, just like in the spreadsheet model, so where is the improvement? The improvement is that the errors are reported to the users as they make the errors and then can correct it on the spot. If I have a giant pile of spreadsheets, the people who typed them are nowhere in sight and I am trying to load them and I get 10,000 errors. What do I do now? The answer is rewrite the terribly broken application. Like a butter knife should not be used as a screwdriver and a shoe should not be used as a hammer and a coffee can should not be used as a hardhat -- similarly a spreadsheet should not be used as a database. A spreadsheet is good to crunch numbers but it has no safeguards to ensure that the data you give it is valid (or at least not nearly the sort that are available with a database). For most real business applications, spreadsheets should really not be the data repositories that they are typically used for. All modern spreadsheets have a fascility to connect to a database and that is where the spreadsheet data should really reside, for a number of reasons (not just the data integrity issue). > A spreadsheet is easier to use than a text >>editor but it is not safer. > >What is unsafe in a text editor? Suppose that you have a file of 12,321 chess players, built up over 10 years time by dozens of typists working away. Now, I am a bit careless and come along and enter 10 new entries. Some of the entries are duplicates. Some of the entries are typographical errors. I just did thousands and thousands and thousands of dollars in damage. >Note that it is possible to save a backup of the data if there is a risk to lose >information. There is a risk to lose information and there is a risk to degrade information and there is a risk to destroy information. Because it is a text file, you will not have any idea with any of those risks materialize. Nothing of great value should be stored in a text file. Data (like a nation's chess player facts) belongs in a database and not in a text file and not in a spreadsheet. >I also think that it may be possible to do everything by C program that read >text files and I think that it may be better solution than the situation today >because it may be possible to change data simply by editing text files when I >read that the situation today is that in case that there is a mistake people >need to delete everything and enter the data again instead of simply fixing the >mistakes. For small scale solutions, C programs and text files can be made to work. But a database is a much better idea. However, the database is only a valuable tool in the hands of a database expert. A badly designed database is no better than a text file or a spreadsheet. Probably, there are some Isralei database experts who will donate their time to help. There is a good start already, but as you have already discovered there are data problems developing in your data and the data is difficult to understand because there is not a clear picture of the data model available.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.