Comments by "Lawrence D’Oliveiro" (@lawrencedoliveiro9104) on "Microsoft Excel just got Python" video.

  1. Remember, Microsoft is not going to offer you Python-in-Excel for free.
    8
  2. Just leave out the “Excel” part.
    7
  3. Python, like Linux, became popular entirely through its own merits, not because there was some marketing-behemoth megacorp promoting the hell out of it. All the marketing-behemoth megacorps have done now is notice that these stars are rising while theirs are waning, so they are trying to embrace the new things just in the hope that some of the shine will rub off on them.
    6
  4. They want to make money from it.
    5
  5. I think a Jupyter notebook would be far more flexible: each cell can hold as much or as little code as you like, and can produce multiple associated items of rich output — graphics, sound, video, maths, formatted HTML, plus interactive widgets to control same — interspersed with cells containing narrative text so your notebook can be self-documenting. And you can include code in other languages besides Python.
    5
  6. They tried, with “Linux is a cancer”, back in Ballmer days. And failed. And they had much more control over the market back then than now.
    4
  7. Think of this as Python+Jupyter+Pandas+etc embracing and extending Excel. Guess what happens next ...
    4
  8. Excel spreadsheets are notoriously hard to debug. What’s worse is that most of its userbase has no idea about debugging techniques anyway, since that requires programming knowledge, and they use Excel because they hate programming. There have been many embarrassing cases of wrong data being published because of Excel spreadsheet bugs -- even in scientific research papers. When other researchers want to replicate those results, they ask for the original spreadsheets, only to discover that their formula ranges are incorrect, that kind of thing. The most depressing case I heard of is that geneticists are changing the names of some genes, because when they import info into Excel, it wants to interpret those names as dates.
    3
  9. Remember that Microsoft is doing this because it has to, not out of the goodness of its heart. The only reason it has to pay attention to Python is because it is dominating the professional-quality data-analytics industry, and leaving the noddy Excel jocks in the dust. This is a way to make those Excel jocks feel they are still somehow relevant, and also pay Microsoft for the privilege.
    2
  10. Or, better still, get rid of the Microsoft part and run the Python+NumPy+Jupyter+Pandas+Matplotlib+etc part directly. For free.
    2
  11. This is an opportunity for Microsoft to squeeze a bit of revenue out of those too dumb to realize that they can drop the Excel part—indeed, the entire Microsoft part—and run this entire advanced Python-based analytics stack for free.
    2
  12. Nobody uses Dotnet for anything important. Microsoft itself won’t use it in anything worthwhile, like Office. Even when they were developing VSCode, they decided to go for, of all things, Electron with JavaScript/TypeScript, instead of DotNet.
    2
  13.  @GSBarlev  He also gave up his post as Python BDFL. Coincidence? You be the judge.
    2
  14. Remember that Microsoft is doing this because it has to, not out of the goodness of its heart. The only reason it has to pay attention to Python is because it is dominating the professional-quality data-analytics industry, and leaving the noddy Excel jocks in the dust. This is a way to make those Excel jocks feel they are still somehow relevant, and also pay Microsoft for the privilege.
    1
  15. It's integrated in lots of open-source apps, like Blender, Inkscape, GIMP ... basically, it’s become very popular glue for content-creation industries, among others.
    1
  16.  @fred.flintstone4099  If it can be integrated into GPL’d software, it can be integrated into anything (open-source).
    1
  17. I don’t use Anaconda. Is that a Windows thing?
    1
  18. Think of this as Python+Jupyter+Pandas+etc embracing and extending Excel. Guess what happens next ...
    1
  19. Hint: everything that Excel does can be done much more nicely by those other tools.
    1
  20. At least you know about LibreOffice, instead of hoary old OpenOffice.
    1
  21. Yes, continue with Python. Think of the Excel jocks who get excited by this, until they discover that, to take full advantage of it, they have to learn Python. Or hire someone who knows it. That’s where you come in.
    1
  22. Or try a Jupyter notebook. Each cell can have any number of lines of code, and produce any number of items of associated rich output—graphics, charts, sound, video, mathematical formulas, even arbitrary HTML. And interactive widgets to let the user control that cell output. You can intersperse code cells with narrative text cells, so the notebook becomes self-documenting. And you can use languages besides Python, if you want. Where is there a spreadsheet that comes close?
    1
  23. Of course. Then when some Excel jock discovers they can’t use this new feature without knowing Python, they’ll hire you.
    1
  24. I still think that.
    1
  25. Now it can have the full-grown sheepdas as well.
    1
  26. It already does. All you have to do is leave out the Microsoft part.
    1
  27. Once they realize that everything they do in Excel can be done better in Python+NumPy+Jupyter+Pandas+Matplotlib+all the rest of it, then Excel just becomes the shell you start to get to the other stuff. Microsoft will likely make some money off the crowd who are too dumb to realize that they can dump the Excel part and use the rest of it ... for free.
    1
  28. It’s a way to squeeze some more revenue out of those Excel jocks too dumb to realize that they can leave out the Excel part—indeed, the entire Microsoft part—and run a full professional-quality Python-based data-analytics stack for free.
    1
  29. They know where the revenue opportunity lies. And that is with Office 365, not Windows.
    1
  30. I used to use Perl, before discovering Python. With Perl, I always felt that there were esoteric implications to the constructs I was using, that the Perl masters understood and I did not. With Python, I found it quite clear from the beginning what all the constructs did. Surprises were few.
    1
  31.  @adamestrada7610  Still only Python 3.7!? Because the current oldest upstream-supported version of Python is 3.8! Are Microsoft providing their own support for 3.7? (Somehow I doubt it.)
    1
  32. Bloomberg Finance did a presentation at a Python conference a few years ago, where they demonstrated how they create quite amazing-looking custom dashboards for their users inside Jupyter notebooks. They even open-sourced the add-on toolkit they use to do it.
    1
  33. Embittered Visual Basic fan speaking?
    1
  34. LO Writer can open multi-hundred-page documents without choking.
    1
  35. Sure, the programmers will use the Python and the Jupyter and the Pandas and all the rest of the data-analytics toolkits they already do. Just leave out the Excel part.
    1
  36. It’s for those Excel jocks who are too dumb to realize you can access all this stuff directly, outside Microsoft’s walled garden, for free.
    1
  37. Guido also gave up his post as Python BDFL. Coincidence? You be the judge.
    1
  38. They did: it was called “IronPython”, built on DotNet. Nobody wanted to use it. Because ... DotNet.
    1
  39. I think also GDPR discourages a lot of 🇪🇺 developers and businesses from jumping wholesale into 🇺🇸-based cloud services.
    1