topic proposal: SPEED

Forum for discussion about the documentation project.
Post Reply
speedfixer
Posts: 606
Joined: Nov 28, 2012 1:27
Location: CA, USA moving to WA, USA
Contact:

topic proposal: SPEED

Post by speedfixer »

Sooner or later, everyone wants their program to execute faster. And, perception of language speed does influence choice of language for a project. WE know FB is fast. A beginner or even a good programmer in another language may not appreciate how fast FB really can be. (Someone: coding more tasks at Rosettacode and/or benchmarks game would help.)

There are so many 'tips' that can be given that it would be a challenge to collect them all in an organized manner.

Many things that an experienced programmer does automatically are mentioned only when someone's code fails to include them, for example.
Some things we assume our compiler system will do for us - does not always happen.
Some require not a small effort, or may even a different philosophy toward the approach to a problem.
Most tips would be valuable in any language.

There should be groups:

basic/general/simple - example: use an integer type when possible
intermediate - example: pass reference to large object by reference instead of by value (not always automatic), avoid strings
advanced/complex - example: loop unrolling, division of number by reciprocal multiplication, threads
fancy/exotic/app specific - example: graphics
----------------------------------------------------------

** Also, I think this (SPEED), as well as Tutorials, should be seperate, individual forum topics, under that large grouping of MISC.**

Tutorials should be considered a part of documentation, but the size and discussion needed to refine one tutorial kind of overwhelms the intent of the Documention forum topic, to my mind. I don't think there should be confusion on where to post a doc question. And I think that the importance of actually correcting 'official' documents should not be watered down with a lot of other noise in the same forum topic. This way, all discussion would be 'offline' to the Documents section. DOCUMENTS could have a sticky reference to each index in the other forums.

A sticky at the top of each new forum topic would be the index to a 'completed' topic list, as proposed already. When discussion slows down enough, it is ready to be indexed.

Any 'finished' specific topic document in that forum would be discussed as would any usual forum topic, and a topic owner would edit/update that topic that help from forum discussion. Hopefully, a topic 'owner' and other contributors would be kind and not feel or use criticism as personal attacks.

david

------------------------------------------------------------------------------------------------------------------------------------
A lot of people - both sons, for example - say I talk too much when I start talking. I can't deny this. It is true.
My parents taught their children 2 things:
If you don't have something to contribute: don't talk. (This works very well.)
If you object - with a good reason - don't be afraid to speak. (This gets all of us into trouble very frequently. Eventually, we either get out or take over!)
------------------------------------------------------------------------------------------------------------------------------------
I make these proposals, without offer to help, for a reason.
I am a problem solver. I do this very, very well.
I am also a father; I think I did a good job. That means I am some form of teacher.
OK, also, before I retired, I toured teaching, and wrote tech and procedural docs.
I have my style - some people don't like it.
I have NO formal programming schooling.
I did have a pro custom programming business for many years: I was drafted into it. I ONLY wrote stuff you could not buy off the shelf - a requirement of hiring me. For me, these were just problems to solve. (A long story why I quit - involved a divorce.)

If you ever answer me with only a quote from a textbook, I will rip you apart.
If you tell me you tried it and it didn't work, we can work on it forever.
If it hasn't been tested to BE true, then it is just an opinion or speculation.
I spent the last 15 years of my career proving to other people that (something in) their reference books and what they were taught was wrong. My job was to fix things where others had failed AT LEAST 5 times before it was allowed to be brought to my desk. That was my job. My company was VERY big. I was very busy. (I loved my job!)

If I write something, and you object - be prepared to defend with working code, not someone's opinion.

I would LIKE to work on some of these tutorials and docs. And would invite LOTS and lots of objections/corrections/inclusions so I can get it right. Or at least acceptable. (I'm retired. I get bored with my personal project, sometimes. I have time.)
Should I try?

I would prefer an assignment, not a self-chosen topic.
I'm not an expert at anything EXCEPT solving problems. Any problem - any area - any field - tech, social, legal, semantic, procedural, ...
Assign a topic. I will research. I will submit a paper. You all will criticise and suggest. I will rework it. Oh, and I will learn about the topic.

Summary:
help FB: more FB in rosettacode and the benchmarks game
new topic: SPEED
new forum topics: SPEED and TUTORIALS
should I try to help write some docs? vetoes, and votes for? (I won't be offended)
Lost Zergling
Posts: 534
Joined: Dec 02, 2011 22:51
Location: France

Re: topic proposal: SPEED

Post by Lost Zergling »

Hello. Like any user, I use this forum and I see a great tolerance in the subjects, which is very significant. I maintained a post in the documentation / tutorial on LZLE (duplicate with a post in General Topic), I think there is a logic to leave it, but I can not give you the assurance. However, I must say that any help in terms of documentation, criticism or suggestions around LZLE on tests of features / volumetrics / speed, etc. would be welcome for me. Thus, other users will probably appreciate a different view of mine. The subject of speed is discussed frequently but it is often inseparable from the functionalities the link between the two being design and optimization, which implies impacts on the structure of the language. I recommend these topics: https://freebasic.net/forum/viewtopic.p ... 3&start=45, https://freebasic.net/forum/viewtopic.p ... 3&start=60. Look out at Marcov's links. Lzle project is small, poor scientifically, unofficial and has a small assumed parasite side but it is constantly evolving, so if the subject attracts you and you would like to participate, be welcome. For the doc: https://freebasic.net/forum/viewtopic.php?f=9&t=26551, for evolutions / corrections / reviews: https://freebasic.net/forum/viewtopic.php?f=8&t= 26533, for the discussion: https://freebasic.net/forum/viewtopic.php?f=2&t=26637
Post Reply