Happy new year 2019
44 views
Skip to first unread message
Matthew Lai
Jan 1, 2019, 12:22:52 PM1/1/19
to Flow Based Programming
Wish all the fbp folks and their family a very prosperous 2019 ahead :)
Any FBP specific plans/ideas/implementations that you ladies & gents want to share here? It's not too late to nail down those resolutions as there are still 364 days to work on them...
My FBP resolutions:
(1) replace socket-based outport<-->inport connection to use "ring buffer in shared memory" to support the bounded FIFO
(2) use Tcl coroutine as the main running loop for each node. Sort of a major change as I'll be switching from event based processing to active message polling (poll the shared memory segments). It does allow a less awkward support for selective port reading as compared to what I use now (a separate traffic cop component) in the event processing model.
The above may or may not play out, but at least I have something to shoot for in the next few months :)
Yours,
Any FBP specific plans/ideas/implementations that you ladies & gents want to share here? It's not too late to nail down those resolutions as there are still 364 days to work on them...
My FBP resolutions:
(1) replace socket-based outport<-->inport connection to use "ring buffer in shared memory" to support the bounded FIFO
(2) use Tcl coroutine as the main running loop for each node. Sort of a major change as I'll be switching from event based processing to active message polling (poll the shared memory segments). It does allow a less awkward support for selective port reading as compared to what I use now (a separate traffic cop component) in the event processing model.
The above may or may not play out, but at least I have something to shoot for in the next few months :)
Yours,
Matt
ern0
Jan 1, 2019, 2:26:50 PM1/1/19
to Flow Based Programming
(I don't remember if I told you, so probably this mail is a kinda duplication.)
FBP (or dataflow programming) is often the _technique_ which is used
in End-User Programming systems. As usual, Wikipedia article is great
in this topic: https://en.wikipedia.org/wiki/End-user_development
I think EUD/EUP is a very interesting area with lot to do. I've found
several videos, presentations in this topic, but still we're on the
beginning of the road. The world needs more and more programmers, and
even if the education is great (it is not), even the programming tools
are easier to use (not), and even platforms are cleaner, easier to
understand (they are not), there is not enough programmers. The USA
and rich countries can get work force by importing people from other
countries, but still not enough.
We will have robots soon, and no one will program them.
My $.2 is that we should build systems which enables end-user
programming, and these kind of systems will beat ones which are only
parametrizable, because requires less professional programmers.
Also, HNY :)
> --
> You received this message because you are subscribed to the Google Groups "Flow Based Programming" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to flow-based-progra...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
--
ern0
dataflow evangelist
FBP (or dataflow programming) is often the _technique_ which is used
in End-User Programming systems. As usual, Wikipedia article is great
in this topic: https://en.wikipedia.org/wiki/End-user_development
I think EUD/EUP is a very interesting area with lot to do. I've found
several videos, presentations in this topic, but still we're on the
beginning of the road. The world needs more and more programmers, and
even if the education is great (it is not), even the programming tools
are easier to use (not), and even platforms are cleaner, easier to
understand (they are not), there is not enough programmers. The USA
and rich countries can get work force by importing people from other
countries, but still not enough.
We will have robots soon, and no one will program them.
My $.2 is that we should build systems which enables end-user
programming, and these kind of systems will beat ones which are only
parametrizable, because requires less professional programmers.
Also, HNY :)
> You received this message because you are subscribed to the Google Groups "Flow Based Programming" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to flow-based-progra...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
--
ern0
dataflow evangelist
Matthew Lai
Dec 31, 2019, 10:29:43 PM12/31/19
to Flow Based Programming
Wish all the fbp (traditional or not) folks and their families a very prosperous decade and 2020 ahead :)
Any FBP specific plans for the coming year/decade? Surprisingly I hit my two previous goals just days before the end of 2019, so writing down your goals does "work"!
My FBP resolutions with 20/20 vision:
(1) Design and implement a motley test crew (Linux native, Linux VM, Mac OSX 10.14, RasPi 2/3/4/zero, UdooNeo) to support regular and regression tests for fbp application that spans across multiple machines.
(2) Going international: setup up FBP nodes in Toronto, Hong Kong, Singapore and Houston to attempt cross country FBP processing. Not sure whether it's practical to so but the thought of it is funky enough to turn me "on" !
(3) Integrate the motley test crew with Bluetooth mesh network (comprises Nordic Semiconductor's nRF52832/833/840, nRF5340 (dual Cortex M33) and nRF9160 (Cortex M33 with radio that supports NBIoT/LTE-M connectivity thru the 4/5G network). RasPi 3/4/zero comes with wifi/Bluetooth combo so the BLE (Bluetooth low energy) connection with the Nordic chips should be feasible.
(4) Upgrade the existing socket-based FBP apps (web scraper for financial data, bigEnglish to automate English to Chinese word/phrase translations) to shared memory and coroutine based FBP version.
(1) Design and implement a motley test crew (Linux native, Linux VM, Mac OSX 10.14, RasPi 2/3/4/zero, UdooNeo) to support regular and regression tests for fbp application that spans across multiple machines.
(2) Going international: setup up FBP nodes in Toronto, Hong Kong, Singapore and Houston to attempt cross country FBP processing. Not sure whether it's practical to so but the thought of it is funky enough to turn me "on" !
(3) Integrate the motley test crew with Bluetooth mesh network (comprises Nordic Semiconductor's nRF52832/833/840, nRF5340 (dual Cortex M33) and nRF9160 (Cortex M33 with radio that supports NBIoT/LTE-M connectivity thru the 4/5G network). RasPi 3/4/zero comes with wifi/Bluetooth combo so the BLE (Bluetooth low energy) connection with the Nordic chips should be feasible.
(4) Upgrade the existing socket-based FBP apps (web scraper for financial data, bigEnglish to automate English to Chinese word/phrase translations) to shared memory and coroutine based FBP version.
This is quite an ambitious list, and I'll be quite surprise if I can finish them all in 2020. Fingers crossing :)
Yours,
Matt
Paul Morrison
Jan 1, 2020, 5:17:18 PM1/1/20
to flow-based-...@googlegroups.com
Hi Matt,
That sounds like a fantastic program! Hope you keep us posted on your progress along the way!
Happy new year to you, and to all the other FBP aficionados out there!
All the best,
Paul M.
--
You received this message because you are subscribed to the Google Groups "Flow Based Programming" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flow-based-progra...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flow-based-programming/f92a0873-cb2a-4a3a-9a9e-fb49f02be14b%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages