For myself and my colleagues, who are also civil engineers and geologists, I have to disagree. The notebook provides the top layer, documentation, and the helper scripts, which live on a shared drive with the notebook, are easily accessible and reviewed by my peers.
-
-
Notebooks have been nothing short of revolutionary for how we complete, communicate, and review our scientific and engineering analyses. "Notebooks aren't good for software engineering" completely misses the point and misunderstand/misrepresents the target audience and workflows
0 proslijeđenih tweetova 5 korisnika označava da im se sviđa -
I won't comment on notebooks for software development, because that's not what I do. I do methods development in computational science, so I need to document methods, both by direct explanation and by examples.
1 reply 0 proslijeđenih tweetova 0 korisnika označava da im se sviđa -
That requires explaining multiple levels of code, which I cannot do with notebooks. A notebook is a single narrative, I need multiple narratives on a shared base of code and data. I want to say "click here for an explanation of the algorithm", pointing to another narrative.
1 reply 1 proslijeđeni tweet 2 korisnika označavaju da im se sviđa -
To sum it up, I want a computational Wiki, not a single page. And I want all the underlying code to be just one click away, without forcing readers to switch to a different tool.
1 reply 5 proslijeđenih tweetova 10 korisnika označava da im se sviđa -
The sad part is that the technology to do this was around in the Lisp machines of the 1980s, and most of it also in Smalltalk environments of the time. Emacs has much of it as well, but crucially lacks graphics.
1 reply 3 proslijeđena tweeta 4 korisnika označavaju da im se sviđa -
Today we have monstrous multi-technology assemblies such as
@ProjectJupyter that are inferior in terms of features and user interface quality, and we celebrate them as great progress.1 reply 0 proslijeđenih tweetova 5 korisnika označava da im se sviđa -
Compare to
@feenkcom's#GToolkit which is a much better design, and done by a small team. But a small team standing on the shoulders of giants: 40 years of Smalltalk. Notebooks are quick hacks that ignore much of the rich experience of the past.1 reply 4 proslijeđena tweeta 8 korisnika označava da im se sviđa -
Indeed, it is unlikely that we would have been able to think of
#gtoolkit in another environment. Lisp could have been the other candidate. The conceptual and pragmatic work done before us around Pharo, and Smalltalk in general, created a world where one can dream differently.1 reply 2 proslijeđena tweeta 6 korisnika označava da im se sviđa -
Interestingly, we wanted to create a new kind of an environment for programmers, and only later realized that we inadvertently also provide an alternative perspective to what notebooks can be. This serendipity is a direct result of working with beautiful abstractions.
1 reply 2 proslijeđena tweeta 6 korisnika označava da im se sviđa
Information theory tells us that when it comes to information, there are no real boundaries. The silos that get created, such as notebook vs code, are typically arbitrary and mostly driven by some historic pragmatic assumptions that are neither conceptual nor necessary.
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.