Anecdotes of how we have made systems and programs, their mistakes and things learned. Real facts and opinions. Algorithms and Problems. Tips and Tricks.
At the beginning of the 70s we had a computer with 32 Kilowords and 12 bits in length each word and to program it encoded the machine language by hand and could also use a posfix notation language called LEAL although it was difficult for us anyway to translate algorithms especially if what prevailed was the calculation, so frequent in the Neuroscience applications that we developed supported by statistics, including multivariate ones.
In those years and also in the 60's, it was common for scientific publications of any level and on any subject to be accompanied by the full text of the program that implemented the solution described in the articles, including many new algorithms.
As a department of Neurosciences that we were and influenced by the great scientists of international stature who trained us, worked with us, or came to see us regularly, we were steeped in the latest in a varied range of scientific disciplines since the study of the brain requires a multidisciplinary approach, which is currently applied to almost all human activity.
Then we fell into an article with its corresponding FORTRAN program that implemented the Simple Link Cluster Analysis, as it is better known in the world.
But it was in FORTRAN! and we still did not have the possibility to use this language, so I took on the task of implementing it in LEAL, for which I had to do a detailed analysis of what was written in FORTRAN, understand it, and then write in LEAL what was relevant. Believe me, it is more difficult than translating from English to Chinese for someone who is not a native English speaker and does not even have a proper dictionary.
Well, in the end I got the translation, but to my perplexity the results of the resulting graph did not match the article, as the article had, as usual, a test data set with the printed solution.
I spent days and days thinking about FORTRAN and LEAL and nothing, I even came to think that the article was a scam and that FORTRAN was the one with the error.
So I approached Dr. Pedro Valdés, I raised the situation and have you seen how in the Sherlock Holmes films he concentrates smoking a pipe looking at something? Well, that was the impression on me that watching Pedro analyze the list of the two programs and to my surprise after no more than 10 minutes he told me with aplomb: "change this for this" and voila! I did everything went as planned !!.
We have been using that particular program for years and it was even one of the cornerstones of the Ph.D. thesis that Peter presented, in addition to contributing multiple results to the department's research.
I will never forget Peter's genius in solving this problem ... an example of the spark that lights the flame of knowledge.
Octavio Báez Hidalgo.