Comments by "Lawrence D’Oliveiro" (@lawrencedoliveiro9104) on "Thu Vu" channel.

  1. Jupyter does rich output and interactive widgets. Quite handy for teaching/exploration purposes.
    22
  2. Jupyter is Open Source, VS Code is not. Also, the overhead of Jupyter is a tiny fraction of that of VS Code. It’s small enough I can leave Jupyter running on my machine all the time.
    3
  3. An IDE is more for developing complete apps. Jupyter is more for exploration of programs and data. If a cell generates audio or video output, Jupyter will automatically provide controller widgets so the user can control playback of that output.
    2
  4.  @jcbritobr  Jupyter gives each cell its own associated output/interaction area next to the cell, rather than forcing all output into a single area.
    2
  5. 3:44 I use a scratch notebook for this. But really, the whole concept of a notebook is about scratchpad programming and exploration.
    2
  6. “Crushing Excel” isn’t really saying much, though.
    2
  7. Jupyter does rich output and interactive widgets. And you can run your code incrementally, cell by cell.
    1
  8. Sounds like Microsoft is trying really hard with its Jupyter-wannabe. Nevertheless, I think I will continue to stick with the real thing.
    1
  9.  @edweard  Just think of Microsoft’s track record for software quality.
    1
  10.  @edweard  And also the question of whether Microsoft will still be putting resources into it a year or two from now, or whether it will have abandoned it in favour of yet another new idea, just like so many of its past abandoned projects.
    1
  11.  @edweard  There has already been one attempt to remove a feature (“Hot Reload”) from VS Code that was there already, only rolled back after widespread protest.
    1
  12.  @edweard  For now, anyway. Look at what happened to PowerShell.
    1
  13. The PowerShell issue I was referring to was regarding their Linux port. Someone filed a bug report pointing out that their built-in implementations of “curl” and “wget” were broken, and worse than that, were blocking access to the full-function Linux versions. The bug report was rejected, on the grounds that they had to maintain backward compatibility with Windows brain-damage.
    1
  14. Put a print expression in a cell.
    1
  15. Jupyter does rich output and interactive widgets. It also lets you run your code incrementally, cell by cell.
    1
  16. Interaction in notebooks is very useful. You know the old saying: “a picture is worth a thousand words”. Well, a picture that moves, that the user can interact with, is probably worth a hundred times that.
    1
  17. Remember what Jupyter is for: it’s for exploration, teaching and “scratchpad” programming. It has these advanced interaction facilities, also including the Notebook API and asyncio support. It’s not an IDE, and it’s certainly not a bloated, inefficient one.
    1
  18.  @nunezkant  It’s still not a tool for advanced data analytics, though. There’s a video by Bloomberg Finance, showing how they build custom analysis dashboards in Jupyter for their own experts. Go learn from them.
    1
  19.  @nunezkant  And Jupyter also runs the asyncio event loop. That means you can have concurrent tasks managing “live” output on different cells at once. See what I mean about a “dashboard”?
    1
  20. Did you know that Jupyter notebooks, those .ipynb files, are just text JSON files? I put them in Git version control all the time. One tip: clear all cell output before committing. That will get rid of the junk. And the result is, you will have nice commits that you can compare cleanly to see what has changed.
    1
  21. Jupyter has interactive widgets and rich output. It also has a Notebook API that you can access from Python.
    1