Hi all, Francesco's words resonated a lot with me, and I felt like telling my own story in case it is of any interest.
I started my MSc in Aerospace Engineering in 2009 and got into open source math software trough Sage around 2010 (probably googling "free alternative to Mathematica"). Actually I did some experiments with the notebook and explored the idea of implementing it using websockets[1]. At some point though, I was discouraged by the difficulties of installing Sage and the syntax differences from pure Python, so I started exploring NumPy, SciPy, matplotlib and SymPy outside of Sage. These libraries essentially covered everything I needed, were infinitely easier to install and the community was bigger - hence I had lots of tutorials and StackOverflow answers. In the meanwhile, I have been organizing the Python Conference in Spain (PyConES) since 2013 and every year there is an "Introduction to Sage" talk which essentially presents extremely basic functionality already available in IPython/Jupyter with plain NumPy and matplotlib. And I find myself thinking "Why do they advertise this instead of the true selling point of Sage: pure math?" I've been giving talks and courses in Europe and Latin America and find myself happy without Sage. I feel sorry! The situation got better over the years (the IPython notebook was born, then Anaconda came to scene so I never ever had to recompile NumPy again) and it's been ages since I last used (and recommended) Sage for any real work. Perhaps Sage symbolics are better than SymPy - but what's the cost? I do not know a single classmate who would find in Sage a key functionality missing in the "SciPy stack". Let me stress again that we are engineers - most of what we do is solving ODEs, plotting and reading from files. I have been reading very closely all William posts about the topic and am honestly worried about Sage's future, because it's a project I still love. Its developers have poured a tremendously big amount of top-quality, unpaid work. But I agree with others that the project diversified too much and maybe should start *dropping* things instead of *adding* things. What about "Being a viable alternative to Magma" - and leave the rest of the Ma's out? What about removing something of those 800 Mb? Splitting core functionality and making it available through pip and conda on PyPI? Stating it as a federation of projects? Focus on its strengths and its present community (that is: pure mathematics) rather than pleasing everybody? Sage's current mission statement is epic and desirable, but maybe it is too broad. My two cents. Juan Luis Cano [1] https://groups.google.com/d/msg/sage-notebook/JbJSULEX3hA/6N73HltFFEEJ On Tuesday, September 29, 2015 at 4:44:07 PM UTC+2, bluescarni wrote: > > On 28 September 2015 at 19:37, William Stein <wst...@gmail.com > <javascript:>> wrote: >> >> 1. Magma is also an Ma. Magma's incredibly good at pure mathematics. >> You seem to be leaving out Magma above. >> > > I admit I know basically nothing about Magma (I did not know it even > existed before joining this list :). > > >> 2. You say "... better served in the Python space by the >> Numpy/SciPy/SymPy/Matplotlib stack as an alternative to the Ma's >> rather than SAGE." Sage includes "Numpy/SciPy/SymPy/Matplotlib", so >> we don't have to worry about that chunk of people with respect to our >> mission statement. >> > > I can install the NSPM stack on any modern platform (Windows included) > with pip (or my distro's package manger, if I am on linux) in probably less > than 5 minutes with a reasonable internet connection. I don't have to worry > about sizeable downloads, virtual machines, containers, emulation layers > a-la cygwin, installing a separate compiler toolchain/python version/set of > libraries, or anything of the sort. > > It is true that you can use the NSPM stack from SAGE, but what are the key > advantages of doing so? It is a honest question, maybe there's something I > am overlooking. > > 3. There is a lot more to mathematics than just what Magma does and >> *also* much more to it than just what Numpy/SciPy/SymPy/Matplotlib do. >> There's a huge amount of interesting things that could be >> systematically computed with in mathematics that no existing package >> does yet. >> > > Yes, nobody has the monopoly on what "mathematics on a computer" means :) > > My comment was merely a marketing/strategic one: I think there exists a > disconnect between the mission statement and what SAGE actually is (and > maybe what the SAGE community wants it to be). > > Cheers, > > Francesco. > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
