Every Developer is Welcome: open sources and development environment by Ian Choi- Lightup Conference
Nov 16, 2023
Conference Website: https://www.2020twenty.net/lightup
#lightup #2020twenty
Show More Show Less View Video Transcript
0:27
which is in support of UNICEF
0:30
We all welcome you into this event. Thank you. Thanks a lot for the introduction
0:36
Hello everyone, my name is Yan Choi. I work in actually Korea in
0:42
APAC region in Microsoft and today then I would like to talk about
0:47
open sources and development environment. So then as you know, the Microsoft now emphasize
0:55
so that I want to bring this topic today to audiences. Before going further, then I would like to mention again
1:06
then please donate if you are very OK, and if you would like to support COVID-19 for many children
1:17
And also then thanks a lot also for lots of sponsors for the event
1:25
event and then this is the feedback page on my speaking and
1:30
also general event feedback. So please to share your feedback. It will be very great
1:38
So then I work in Korea as a part of developer relations
1:44
team and then my main role is to work with many developer
1:49
communities in Korea and also APEC area. For example, like like Singapore, New Zealand, and also Vietnam, Thailand, Indonesia, and Malaysia, and so on
2:00
There are many countries in APEC, so but due to COVID-19, then I could not travel a lot
2:08
Instead, I'm happily joined with this kind of online event. and today then my direct talk by dividing into three different topics
2:21
One is how Microsoft has been has evolving with open source open sources
2:30
and second one is that I direct to some highlight what which kind of open source things were emphasized on last period of 2020
2:38
2020 and third thing is then I would like to demonstrate or show that how
2:44
AI world is developing with open sources and also development environment. And then first of all
2:51
then let's start with Microsoft and open sources. And this is the statistics for recent
3:02
a few years so then still there many which were disclosed on Linux.org
3:38
Linux kernel mailing list, and then on June 2019, then the ratio of Linux operating systems
3:47
on the top of Azure is more than 50%. So it means that now Microsoft really loves Linux
3:56
As Satya said, initially said that Microsoft loves Linux, So which means that Microsoft really
4:33
loves open sources. So then this is some recent progress on Microsoft
4:41
so then actually then before Satya became CEO and then Microsoft also has have had been supporting
4:51
lots of open source technologies like TypeScript and also many other open source technologies using code flex
4:58
codeflex, but however then after satya officially became CEO and then some motions like Microsoft
5:09
loves Linux and also that Microsoft Microsoft really loves open sources and developers such
5:17
motions or more driven. So for example that in 2014, they actually the Microsoft created
5:52
Microsoft product while Dynet core is completely open source. But this is Dynet is still evolving with
6:00
working with open source communities and in 2015 then Visual Studio Code is released
6:07
Visual Studio Code is one of black news from Microsoft because that
6:13
the first of all the Microsoft still. The religious Visual Studio, for example, Visual Studio 2019
6:21
But however, then Visual Studio Code is also being developed using open source technologies
6:28
For example, Visual Studio Code is based on Electron, which is very popular to open source developers
6:35
And also then all the sources are open to GitHub. And everyone can download and see how Visual Studio Code is implemented or that anyone can contribute to enhance Visual Studio Code functionalities
7:19
Then then the core 1.0 was released and also another break news was that Windows subsystem for Linux in Windows 10 were released for developers which means that on Windows still then there are many PowerShell users
7:35
But however, then still today then we might need to admit that relatively many
7:42
Linux there are also relatively many Linux based developers, for example, using shell script
7:48
So then those people are quite familiar with Linux operating systems, while then they are still struggling with executing using command
8:00
commands on CMD console or the power share console. So this is very break news and then Microsoft really would like to collaborate
8:10
with Linux developers and also the Microsoft joined Linux Foundation. Then Microsoft officially is recognized as top open source contributor
8:22
on GitHub and also then there are more more more activities like Azure Kubernetes service based on
8:30
completely open source Kubernetes was launched and also Microsoft officially says that said that
8:36
Microsoft officially contributes to lots of project on top of Kubernetes and also that SQL server on
8:45
were released in 2017 and Windows source code was moved to git based
8:51
And also then in 2018 another break news was that Microsoft announced
8:57
that Microsoft will acquire would acquire GitHub. It is also great news
9:03
and today then more than 5000 Microsoft employees are committing to open source
9:09
projects on GitHub. It is also very good news. So then today then I'm using Windows 10
9:18
Then when I search Linux on Windows App Store, and then I could see lots of many interesting research
9:26
like this is Ubuntu, this is Kali Linux, this is Ubuntu 20
9:32
4 and Ubuntu 18 or Debbie and Suseb Linux or many Linux distributions
9:39
And also then the below picture was been very interesting for me
9:44
Because I first saw this picture in 2016 or 17, which means that Microsoft really would like to
9:53
collaborate with Reddit. So this could come from strong partnership, enterprise partnership between Microsoft and Reddit
10:03
which is very strong and also then the right picture is recent article on May or June
10:11
And then this says that Brad Smith is one of senior leadership
10:16
team members in Microsoft Corp. And then he says that Microsoft were completely wrong about open source in 2000
10:24
So now that Microsoft really loves open sources and also then on GitHub homepage
10:30
you can find customers, many customer stories and also including Microsoft. or source version controls on
11:05
for example, team foundation servers, but they now will move to many
11:11
many repositories are now moved to GitHub based repositories. Visual Studio code is very great
11:19
It is a country the most popular open source ID integrated developer development environment
11:27
So this is the statistics from Stack Overflow last year, And then according to this and then we just to the code is on the top of the most popular development environment and also their own Microsoft
11:41
We is called GitHub repository, then there are many stars and folks also that you can say or share any issues or even you can contribute to your source code and make a playlist
12:48
some for we really require sponsorship. So Microsoft decided to sponsor as keynote sponsor
12:57
so the Microsoft really would like to collaborate with many developer communities
13:02
I really have. I'm really happy as one of employees in Microsoft which who she then Microsoft
13:10
is very working with truly working with developer communities. And also then for the second topic
13:20
let's move to the second topic, and then I would like to highlight some Build 2020 things
13:31
First of all, then Build 2020 was held on last May 19-20
13:40
according to US time zone. And then when you see the poster on the left and right
14:41
May was held very for 48 hours without no break like this lighter summit because
14:52
lighter summit is currently happening for 24 hours which is very great So the opening for 48 hours would mean that many speakings are still
15:06
happening for 48 hours and then Microsoft thought that there can be three different time zones
15:15
One is US based time zone and the second one is Asia based time zone and the third one
15:21
is Europe based time zone. So then some keynote speakings were duplicated for Asia time zones and also Europe time zones
15:33
And then while then some local speakers from Asia region or Europe region also participated in as one of speakers
15:43
which was very great. And also then all the events were connected for totally for 48 hours
15:51
like this light up conference so that it was really great and then let's see then how open
16:01
source activities were emphasized during last build actually then this picture is one of my favorite
16:10
uh scenes during last build 2020 then there have been many open source related activities
16:18
announced about however that this thing is very great for me because then window on the top of
16:26
windows 10 then linux is operating actually then windows subsystem for linux uh we can sometimes
16:33
we can say then wsl windows subsystem for linux and this was first announced on around 2016
16:43
which were about 4 to 5 years ago, and at the time that I was really shocked
16:50
because that on Windows and previously then when I would like to use Linux
16:56
I need needed to install virtual machines on the top of Windows like with using
17:03
Hyper-V or VirtualBox or VMware, and then I need to install Linux
17:08
so then there were performance degradation, and also then some files on the Windows based folders and Linux
17:17
based folders were not compatible so that I needed copy and paste
17:22
to some files and then I need to compile on virtual machines
17:26
using Linux environment and then I need to copy again to Windows
17:30
based environment and so on. But now the Windows subscription for Linux is very great because it is the file
18:07
on Linux and then files permissions are managed using read, write and executable for owners
18:17
groups groups and also other users. So while on Windows based environment
18:24
and then country then many people are using NTFS based file systems
18:29
and then Windows now supports very fine grained permissions. So yeah, it was very great and I'll demonstrate something if someone are still not familiar with Windows subsystem for Linux
18:47
And during last build 2020, then some GUI stops were emphasized. For example, the opening GIMP GIMP application on the top of Windows using Linux subsystem
19:04
and also then windows terminal were announced uh uh officially as 1.0 version so this was also
19:14
really great and then windows terminal uh previously then uh someone are still very familiar with this
19:23
this screen this is a command prompt and then we usually execute this thing using cnd but however
19:33
that no tab no tab is supported and also some there are basic customization for example that
19:40
when i click default properties and then i could change very basic colors so then there are not
19:49
enough functionalities for recent users who are familiar with command line based environment
19:56
especially for developers. But now the Windows terminal is very great. For example, then let me
20:04
show one Windows terminal and then this is current Windows terminal and then I can execute for example
20:13
command prompt and also Windows PowerShell. And then as you see then this screen is black based
20:20
and gray colors for fonts. But however that this screen is based on
20:26
some something different color so I can customize some things by going to settings and also then I can
20:34
customize by creating some more tabs and also that I can also
20:40
integrate with Azure Cloud Shell which works with directly with Azure Cloud which is very great
20:50
OK, and let's move the next thing and then also then on last video 2020 and then GitHub is again emphasized
21:01
So Satya on the keynote Satya emphasized that there are so many developers in 2020 on GitHub
21:11
then there are 50 million developers on GitHub. Those statistics were very higher than last year
21:21
Last year, then Satya mentioned that there were about 36 million developers
21:27
but now Satya says that on GitHub there are 50 million developers
21:35
So more developers are still coming on GitHub, which is very great
21:40
And also there's a team emphasized developer productivity even for remote developers
21:47
So then this is one of interesting thing. So this is code space and the code space means that we just you called can be executed on the top of cloud on Azure
22:00
not only on Azure, but also on the top of GitHub. So then there are currently two different versions
22:06
One is Visual Studio Code Spaces, which is on the top of Azure
22:12
and then you can go to Visual Studio Code Spaces webpage, and then just create a code space virtual machine
22:22
to manage Visual Studio Code on the top of virtual machines And since Visual Studio Code were developed have been developed using Electron and then it surely
22:36
supports web browser-based functionalities. So then CodeSpace is now operating on the
22:43
top of web browsers and so yeah let me demonstrate these CodeSpace
22:49
And then this is Visual Studio code space. So then you might see that this is web browser
23:01
This is a web browser. So while then, let me see. So, okay, I'll execute separately on Visual Studio code
23:16
and this is a my visual studio application, Visual Studio Code application on the top of Windows
23:24
So currently then it is executed on the top of Windows, my laptop while then this Visual Studio Code space
23:32
is running on the top of web browsers. And then it is actually hosted on online
23:40
that visual studio.com. So then how to do that? then you can just go to
23:47
Visual Studio Code Spaces web page and then create an instance. And then you can do seamless development environment
24:00
For example, this kind of comment palette is very well supported on the tab of Visual Studio Code Spaces
24:09
And also even then you can open, For example, click this and then code space using connect code space
24:17
although some extensions would be needed and then you can connect to code space and
24:22
then connect this code space. This is already connected to a code space because I opened using web browser
24:31
And then why then this is another example, so let me reconnect again
24:36
This is an interesting because then the URL is on the top of GitHub.com so
24:47
Which means that this is another version of code spaces supported by GitHub
24:54
so the Git machine's Visual Studio code is open sources and then GitHub
25:30
currently the preview for on the top of GitHub. But however, this is hosted on GitHub
25:37
How to that? And then when you join in GitHub. It's a code space preview and then you might
25:48
click some your preferred repositories, like for example, code space test wrapper
25:54
And then usually then you can clone using this URL, But however, then when you join in the preview program
26:02
then you can click this open code spaces and then you can also create new code space
26:08
But however, I already created one code space. So when I click this and then it loads
26:14
on the top of your web browser. So which means that repositories are many synced
26:22
and using code spaces and you can develop in details on your source code
26:28
and then change and then make a commit or make commit or pull requests
26:34
which is really great. And then even you go to go back to your home or even you are using your iPad or
26:42
Galaxy Tab and then you can continuously develop without changing your development environment context
26:50
For example, then it is very true that when I switch my laptop
26:54
So I'm using both my MacBook Pro plus Surface Book 2. But however, then sometimes I need to bring MacBook
27:04
or the other times that I need to bring Surface Book. So in that case, and then some development environment
27:11
and are actually not synced together. So using this and all the development environment
27:17
should be same. So it is really great to work seamlessly, which is very great
27:24
And so then did you see this is the screenshot on Visual Studio code and then there are so many extensions
27:34
Also, please check more extensions as much as possible. It is really great. And also then there are some
27:45
adjusted web apps were emphasized during last year 2020. It is also really great because then previously
27:54
then when you would like to integrate precast or comments with CI and CD and
28:02
then some built web pages would be needed. So for example, then this static web pages
28:10
would require basic login functionalities. For example, only authenticated users would see
28:16
or then there can be some basic login or basic functionalities to interact with back end
28:24
So now the Azure Static Web Apps are really great because it supports to create
28:32
and or to integrate with static web apps on your public catch or commit
28:38
And on the front end, it will be hosted on the top of Azure Static Web Apps
28:44
While there's some backend functionalities are very synced with Azure Serverless functions
28:50
And then you can easily check for, especially for front end developers
28:55
and it is a little bit difficult for them to see whether
29:01
Uh, local changes are really reflected to cloud, and so they really want to see using a check using web browsers
29:11
In that case, in the state web apps would be really great because it integrates with commits and pull request
29:17
So for any kind of commit or precast, and then the front web page developers
29:24
can easily check using state web apps and then see whether there are still bugs
29:30
or bugs are well resolved, or there are still many bugs or other bugs like this way
29:38
And finally, then, Jandy Core was emphasized on last build, which is very good
29:45
So then on upcoming November, then it is supposed that there will be .NET 5
29:53
Will be a combined version between .NET frame of 4.8 plus .NET core
29:59
three so and then yeah i really uh expect this thing on upcoming november so but uh on last build
30:11
and then danicore 3.1 were emphasized as it is a kind of long-term support edition so that you can
30:21
assure that this version is staying for many years and then there are also many kind of
30:28
other functionalities like self-contained executables and also then modern web uis with
30:34
the asp dynamo laser like this way so then uh next i would like to emphasize how ai world is
30:43
improving with open source and developer environment so this is the recent history for ai world area
30:54
So then when you see, especially for AI world, and you might notice that open source is open sources
31:04
are really inevitable to work with AI domain. For example, then when deep learning first appeared
31:14
and then deep learning was first demonstrated on many research contests. for example, like like resnet or many other
31:27
contests and then Seattle were used as a tab of open sources to
31:33
implement with AI or machine learnings or using deep learning function with this. And now then there are also many open
31:44
sources, but country then issues that TensorFlow and also So Pytorch, two open sources are now very popular and also Keras is very popular
31:57
While then other open sources like initial Theano or Torch were replaced as Pytorch
32:05
but however, Kafe or the other open sources are not too much popular as current popular
32:13
open sources, AI open sources like TensorFlow, Keras, and also Pytorch. And Microsoft also first released CNTK, Computational Network Toolkit, on April 2015
32:27
But also then this was evolved as Microsoft Cognitive Toolkit on January 2016
32:34
But however, currently Microsoft really would like to work with open sources
32:40
Microsoft supports not only TensorFlow, but also Keras and PyTorch on the top of
32:47
Azure using data science virtual machines. And also then Microsoft is now emphasizing
32:53
more on Onyx. Onyx is a project for AI interoperability. For example, there's still then some AI developers
33:01
are using TensorFlow while other AI developers are using PyTorch. They really would like to collaborate together
33:10
together on its onyx has two goals one goal is to resolve ai interoperability using different
33:21
frameworks and or hardware and the second one is that how the performance would be higher using onyx
33:29
so then only microsoft invest uh is contributing more on onyx and then it is now uh one of graduate
33:40
project in Linux Foundation AI. And there are many machine learning and developer
33:48
tools on the top of Azure, but I would like to emphasize now four things
33:53
One is Visual Studio code, which I already explained, and the second one is Azure notebooks
33:59
When you go to notebook.azure.com, then you might see that you don't have to install
34:04
any kind of Jupyter notebook on your local environment. Instead then you can experience Jupyter notebooks on the top of Azure without installation by going to notebook and also then command line would be very important and also the Microsoft
34:23
supports many popular frameworks like Pytosh, TensorFlow, Cyclone or Onyx and then this could be a tab of Microsoft
34:33
powerful infrastructure using by supporting CPUs not only GPUs, but also GPUs or FPGA and also then there are many
34:43
Microsoft based Azure services like not only machine learning virtual machines, but also Azure machine learning services which
34:51
supports very well and on auto ML plus CI CD integration using
34:57
Azure DevOps or and so on. And of course, and then Microsoft
35:03
Support provides domain specific pre trained models for For example, computer visions is very great
35:11
And so those are four screenshots. One is Jupiter notebook so that you
35:19
can go to notebook.azure.com and then you can easily execute Jupiter notebook
35:25
without any installation on your local environment. On the right picture is then when you
35:32
use Visual Studio Code, things on your visual studio code environment and even then it supports to open jupyter notebook
36:10
files and also then if you install jupyter notebook on your local environment and if you specify
36:17
the url or end point for jupyter notebook and then you can connect with jupyter notebook and
36:23
then when you execute on the tab of visual studio code and then it will be executed on the tab of
36:28
of actually Jupyter Notebook. So then using one ID, we just code and you can work 100
36:36
with Jupyter Notebook. And the third one is the screenshot of how
36:42
Microsoft Azure support AIs by supporting lots of many architectures. You might see on some documents on reference
36:54
architectures on docs.microsoft.com for not not only for machine learning, AI, but also other domains
37:01
And finally, a screenshot is actually, I took this screenshot in Korean
37:07
but however, this screenshot is for data science virtual machines. And also on last build
37:16
and then actually then WSL2 were announced. WSL2 is a new version
37:23
So then when I go to this terminal again, And then one interesting thing is that this is on the top of WSL2
37:36
And then, but so one great thing is that this is my, sorry, this is my Docker desktop application
37:49
so I installed Docker desktop and then now they eat directly supports to use WSL2 based engine
37:57
So then for example, then this is my Linux environment, and then for example, then when I execute Docker PS
38:06
and then actually then this comment, when I execute this Docker comment and then the result is
38:12
actually on this Docker desktop application. So the Docker is now hosted on
38:20
Docker at SQL application on my laptop, and then this Docker command is completely
38:27
synced with this Docker desktop environment which means that there have been performance advantages on Docker Not only this and then although it was not implemented on last build on last May
38:44
but however they on June and of this first to say that on last
38:50
May then WSL 2 plus GPU were announced so that not only
38:57
Microsoft pages. For example, this is from Microsoft page as far as I know, and then on Microsoft page and WSL will support GPU computer workflows like NVIDIA CUDA plus DirectML, which is the next versions for DirectX for machine learning developers
39:22
So then those two are announced and at the same time, the NVIDIA announced that some additional functionalities to support WSL2 will be released as preview very soon
39:37
So then when you see this small demo video and then on the top of WSL2, then you can execute machine learning scenario
39:49
And then actually then this is using GPUs on TensorFlow on the top of
39:55
WSL Windows Subsystem for Linux. So that also the as I said
40:03
Nvidia also announced. That this is now public preview, so then you can download and execute and then
40:14
This is the web page. This is the web page on nvidia developer.nvidia.com slash CUDA.WSL
40:26
It explains in very details on the architecture or technologies on how to support
40:37
GPUs on the top of Windows subsystem for Linux. Linux and then you might read more details
40:46
Just one thing I would like to mention is that it is
40:50
country for preview purpose, which means that it requires you to join in Microsoft Windows
40:57
Insider program, which means that your Windows 10 would be converted for
41:04
evaluation purpose for developers and then you can additional install WSL2 related environment to better support
41:17
Nvidia GPUs. So then it will be really great if you try this
41:24
And then currently, just one note, it is still on preview. So performance, there are still issues on performances
41:34
and I hope that this thing will be enhanced very soon. And so and then fine and this is also Microsoft AI
41:46
So there may even then when you go to Azure that Microsoft
41:52
dot com and or by searching Azure Face API on your mobile app
41:58
or mobile web browser and then you can go to this site and
42:03
then you can upload any kind of pictures you by specifying or then you can upload some images with your camera
42:11
for photo camera or photo folder. And then you might see that this is
42:18
perceived emotion recognition. Then happiness is 100% like this way. So then this is achieved using open source technologies
42:31
match with Azure cloud. And then this thing is then how Microsoft thinks regarding AI
42:39
So Microsoft really considers that for AI, then trustworthy is very important So AI should be responsible on machine learning So then this is also built 2020 highlight
42:57
and then Microsoft really would like to support whether your AI model is biased or not
43:04
For example, then now then female and male problems or black and white issues are very serious
43:13
So then we might need to see with you a great unbiased view
43:20
So which is very important. So then Microsoft Azure now supports very well then how to evaluate
43:28
and how to monitor such biased views. And finally, then I would like to mention one additional thing
43:35
So this is on WSL 2 and then for example, This is for Ubuntu 18, and this is for Ubuntu 20
44:16
So then when I execute code, that code space at that
44:26
and then it opens Visual Studio Code on the top of Windows
44:33
And which means that although I'm using WSL2 environment, using this command line. However, then this is open using
44:44
Windows environment and then this is actually on the top of WSL, so which means that when you call debug or execute your own
44:53
command on this and then actually then the execution is happening on the top of WSL, not on the windows. Of course
45:03
then you can switch on. Windows based environment. For example, reopen, reopen for the in Windows
45:11
So for example, then when I go to terminal, then currently then this is
45:16
Linux based terminal environment. But however, when I switch to reopen folder in Windows and then
45:23
it goes to Windows based environment, so now then it is on my command line
45:31
prompt environment using on the top of Windows. So like this way, then Microsoft very would like to collaborate with open sources, including Linux
45:41
user Linux environment, also then GitHub open source developers or and so on
45:48
So Microsoft is very welcome to every developers and then such message were very well emphasized in last build 2020
45:57
So if you are interested in more resources, then please go to mybid.microsoft.com
46:06
to see many interesting build sessions. So then Microsoft is still evolving with Azure
46:15
So Azure is we Microsoft hopes that Azure really supports all of you to invent with your own purpose
46:27
OK, thanks a lot for listening to my presentation and this is
46:31
the end of my presentation. Hope that you are more aware on
46:36
open source sheets on Microsoft. Thanks a lot. Thanks, thanks for providing this wonderful insights on the
46:46
obsession. So thanks for joining in and thanks for taking up this time of your schedule to provide us this insight. Thank you
#Software
#Windows & .NET
#Open Source


