Use of old style random number seeding in documentation

[mikecroucher]

I noticed in the documentation for LXS https://github.com/UniprJRC/FSDA/blob/master/toolbox/regression/LXS.m that you use the old style of seeding for random number generation

randn('state', 123456);

This is a bad idea since it doesn't just seed the RNG, it actually changes the algorithm for random numbers generation to a very old one. Details at https://walkingrandomly.com/?p=2945

I suggest you do the following instead

rng(123456)

[MarcoRianiUNIPR]

Dear Mike, Thank you very much. Inside the .m files which contain the code we have always used rng. In the examples which illustrate the use of the m file (that is inside %{ .... %} it is true that sometimes we have used the old notation with randn. However, you are absolutely right, that this should not be done. In the next days we will replace all instances of randn with rng Thank you and please do not hesitate to contact us for any suggestion you might have Best Marco =============================================================== Marco Riani, PhD Professor of Statistics Department of Economics and Management Division of Statistics and Computing University of Parma (http://www.unipr.it<http://www.unipr.it/>) Via J. Kennedy 6 43100 PARMA ITALY Director of the Interdepartmental Centre Ro.S.A. of the University of Parma (http://rosa.unipr.it<http://rosa.unipr.it/>) PHONE: +39 0521 902473 FAX: +39 0521 902375 e-mail: ***@***.******@***.***> http://www.riani.it<http://www.riani.it/> https://www.researchgate.net/profile/Marco_Riani https://scholar.google.com/citations?user=VDuzaqsAAAAJ&hl=it https://sciprofiles.com/profile/MarcoRiani [https://avatars0.githubusercontent.com/u/55737234?s=400&v=4]<https://github.com/UniprJRC/FSDA/wiki> https://uniprjrc.github.io/FSDA/ -------------------------------------------------------------- Special Issue of Stats: Robust Statistics in Action II https://www.mdpi.com/journal/stats/special_issues/EAYA8LSOPQ Deadline for manuscript submissions: 30 June 2024 Da: Mike Croucher ***@***.***> Inviato: lunedì 12 febbraio 2024 15:02 A: UniprJRC/FSDA ***@***.***> Cc: Subscribed ***@***.***> Oggetto: [UniprJRC/FSDA] Use of old style random number seeding in documentation (Issue #64) I noticed in the documentation for LXS https://github.com/UniprJRC/FSDA/blob/master/toolbox/regression/LXS.m that you use the old style of seeding for random number generation randn('state', 123456); This is a bad idea since it doesn't just seed the RNG, it actually changes the algorithm for random numbers generation to a very old one. Details at https://walkingrandomly.com/?p=2945 I suggest you do the following instead rng(123456) - Reply to this email directly, view it on GitHub<#64>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ANI5YF3WTRRDYDL5YDAC2QDYTIOFZAVCNFSM6AAAAABDEY6GHCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEZTAMRQGYYDMMA>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.******@***.***>>