Sho与numpy / scipy [英] Sho vs. numpy/scipy

问题描述

Hi Sho Team

Hi Sho Team

Sho看起来相当惊人，似乎最终填补了.NET和数值计算之间的空白。我有几个问题：

Sho looks quite amazing and seems to finally fill the gap between .NET and numerical computing. I've got a few questions though:

- Sho如何与numpy / scipy重构项目有关，微软与Enthought合作（http://www.enthought.com/媒体/ SciPyNumPyDotNet.pdf）？在我看来，这些项目有很大的重叠，除了Sho语法更面向Matlab
并且numpy / scipy语法更加面向Python。我在这里完全错了吗？

- how does Sho relate to the numpy/scipy refactoring project, where Microsoft is partnering with Enthought (http://www.enthought.com/media/SciPyNumPyDotNet.pdf)? It seems to me that these projects have a big overlap except that the Sho syntax is more Matlab oriented and the numpy/scipy syntax is more Python oriented. Am I totally wrong here?

- 我也错过了Book of Sho中的函数概述（例如现有的函数，如'repmat'和'reshape'没有提到）

- I am also missing a function overview in the Book of Sho (e.g. existing functions like 'repmat' and 'reshape' are nowhere mentioned)

- 进一步的发展计划呢？我缺少一些标准函数，如unique，sortrows

- what about further development plans? I am missing a few standard functions like unique, sortrows

- 是否有可能在商业上使用该软件？

- will there ever be the possibility to commercially use the software?

谢谢，

Felix

推荐答案

感谢Felix，这些都是很棒的问题;我现在就能给你最好的答案。

Thanks Felix, these are great questions; I'll give you the best answers I can for right now.

1.  numpy / scipy：我的理解是Enthought项目旨在使NumPy和SciPy与IronPython完全兼容，并且可以从IronPython中使用，同时我们考虑到更广泛的.NET受众。 我们的目标是通过任何.NET语言（IronPython，C＃，Managed C ++，F＃等）使用Sho库
可用（和友好）。

1.  numpy/scipy: my understanding is that the Enthought project is geared towards making NumPy and SciPy fully compatible with and usable from IronPython, while we have a broader .NET audience in mind.  Our goal is to have the Sho libraries by usable (and friendly) from any .NET language (IronPython, C#, Managed C++, F#, etc.).

2。你是对的，Sho Book并没有描述每一个功能 - 它更像是一个入门而不是一个完整的命令参考。 您可以从Sho内部获得更全面的报道 - 输入帮助（帮助）将向您显示
浏览命令文档的各种机制。

2. You're quite right, The Book of Sho doesn't describe every function - it's meant more as a primer than a full command reference.  You can get more comprehensive coverage from inside Sho - typing help(help) will show you a variety of mechanisms for browsing command documentation.

3。我们仍然积极致力于Sho，当然欢迎有关新功能的建议！ 对于您所询问的特定功能，有一些简单的方法可以使用Sho。 对于sortrows，如果d是向量，你可以
只做d.Sort（）。 如果它是一个矩阵并且你想按列col排序，你可以链接操作并执行d [（d [：，0] .SortIndex（））[1]，：]。 unique（）也是我经常做的事情，它不是单行的，我们应该为它做一个快捷方式
。 我经常做一个"线"。以下列方式（其中d是整数的向量）：  h = dict（）; [h.Add（elt，0）for elt in d]; res = IntArray.From（h.Keys）。 然后res将包含唯一值。

3. We're still actively working on Sho, and certainly welcome suggestions for new functionality!  In the case of the particular functions you're asking about, there are some easy ways to do it with Sho.  For sortrows, if d is a vector, you can just do d.Sort().  If it's a matrix and you want to sort by column col, you can chain the operations and do d[(d[:,0].SortIndex())[1],:]. unique() is something I do fairly often as well, and it's not a one-liner and we should probably make a shortcut for it.  I often do it one "line" in the following way (where d is a vector of ints):  h = dict(); [h.Add(elt,0) for elt in d]; res = IntArray.From(h.Keys).  res will then contain the unique values.

4。商业可用性 - 很好的问题。请注意，即使对于商业实体，当然也允许使用Sho进行内部原型设计;我们不允许出于商业目的运送这些物品。 这主要是因为Sho仍然是研究原型，
，我们无法提供全面商业产品现在需要的那种支持。 但是，如果我们看到客户对此有强烈需求，那么我们肯定会考虑这一点。 如果您对某项使用有特殊问题
您不确定是否符合许可，请随时向我们发送
行 shofeedback @ microsoft.com 。

4. Commercial availability - great question. Note that internal prototyping with Sho is certainly allowed even for commercial entities; we don't allow shipping the bits for commercial purposes.  This is mainly because Sho is still a research prototype, and we wouldn't be able to provide the kind of support that a full-on commercial product would need right now.  However, if we see strong demand for customers to have this, it's something we will definitely consider.  If you have particular questions about a usage that you're not sure fits the license, feel free to drop us a line at [e-mail] .

我希望能解决你的一些问题！

I hope that clears up some of your questions!

这篇关于Sho与numpy / scipy的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！