Recent changes to wikihttp://sourceforge.net/p/maxima/wiki/2014-07-14T12:34:52ZDiscussion for Orthogonality page2014-07-14T12:34:52Z2014-07-14T12:34:52ZRobert Pollakhttp://sourceforge.net/u/jondo/http://sourceforge.netb4ccf1a3c4d9ab4bc20c6fd53ac33c90706a8ccc<div class="markdown_content"><p>Vectors sometimes are 1-column matrices, sometimes they are lists.</p></div>Comparison modified by Stavros Macrakis2014-07-10T20:16:58Z2014-07-10T20:16:58ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.netbf40c64ee7089de553b6a4e6fcc67bc28f977de1<div class="markdown_content"><p>Some quick hits:</p>
<ul>
<li>
<p>Use realroots for univariate polynomials, e.g., sign(4<em>q^2+4</em>q+3) => pos; sign(4<em>q^2+4</em>q+1) => pn</p>
</li>
<li>
<p>Very basic/simple interval analysis, e.g., sign(sin(x)+cos(x)+2) => pn (interval <span>[-1,1]</span>+<span>[-1,1]</span>+<span>[2,2]</span> => <span>[0,4]</span>)</p>
</li>
</ul></div>Proposed work areas modified by Stavros Macrakis2014-07-08T18:05:51Z2014-07-08T18:05:51ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.net77c24d2b5fdc2909fdb89de53abfa15edec45cf4<div class="markdown_content"><pre>--- v5
+++ v6
@@ -23,3 +23,5 @@
* [Scope] Decide on and implement a consistent, modern scoping system (presumably lexical, but how does that interact with symbols?)
* [Interval arithmetic] Support one or more varieties of interval arithmetic -- with symbolic or numeric bounds.
+
+* [Lists and sets] Full support for a variety of operations; compare R. (This is pretty easy, we just need to develop appropriate function definitions.)
</pre>
</div>Orthogonality modified by Stavros Macrakis2014-07-08T18:04:19Z2014-07-08T18:04:19ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.netac8f2c386db1a1159085d58e0180f5641dde8cfa<div class="markdown_content"><pre>--- v1
+++ v2
@@ -1,4 +1,8 @@
-Here are some cases where Maxima is not orthogonal and could be improved:
+General orthogonality goal: every type of object should work wherever it makes sense.
-* Preserving bfloats in rational expressions (keepbfloat?)
-* Preserving floats and bfloats in matrix expressions
+For example, wherever an approximate number makes sense, both exact numbers (integer and rat) and floats and bfloats should be supported.
+
+Here are some cases where Maxima does not do this:
+
+* bfloats in rational expressions (keepbfloat?)
+* floats and bfloats in many matrix operations
</pre>
</div>Proposed work areas modified by Stavros Macrakis2014-07-08T17:59:20Z2014-07-08T17:59:20ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.neta9e6f3bd1ffe4258ebc8f1b8ed7a625fb437f114<div class="markdown_content"><pre>--- v4
+++ v5
@@ -8,7 +8,7 @@
* [Comparison] Coverage of more simple cases in comparison (e.g., x^2+2*x+2>0 or sin(x)>=-1).
-* [Limits] Improve limit, probably incorporating tlimit into the main algorithm.
+* [Limits] Improve limit, probably replacing the current algorithm with gruntz (and tlimit?).
* [Formatting expressions] Review the various existing ways of reorganizing expressions -- in particular, contrib/format should probably be upgraded to share after we review its functionality, test suite, etc. Beyond that, can we unify semantically or at least document the various ways of manipulating expressions?
@@ -21,3 +21,5 @@
* [Arrays] Simplify and unify the various array/list systems.
* [Scope] Decide on and implement a consistent, modern scoping system (presumably lexical, but how does that interact with symbols?)
+
+* [Interval arithmetic] Support one or more varieties of interval arithmetic -- with symbolic or numeric bounds.
</pre>
</div>Proposed work areas modified by Stavros Macrakis2014-07-08T00:37:46Z2014-07-08T00:37:46ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.net00e67a42b41a177013312d06d3faa56fe466c9cd<div class="markdown_content"><pre>--- v3
+++ v4
@@ -19,3 +19,5 @@
* [Symbol and function properties] More general and more consistent handling of properties.
* [Arrays] Simplify and unify the various array/list systems.
+
+* [Scope] Decide on and implement a consistent, modern scoping system (presumably lexical, but how does that interact with symbols?)
</pre>
</div>Proposed work areas modified by Stavros Macrakis2014-07-07T23:33:11Z2014-07-07T23:33:11ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.net85c8b7de93ba349b872c458d36db9c008f8a31e8<div class="markdown_content"><pre>--- v2
+++ v3
@@ -18,3 +18,4 @@
* [Symbol and function properties] More general and more consistent handling of properties.
+* [Arrays] Simplify and unify the various array/list systems.
</pre>
</div>Orthogonality modified by Stavros Macrakis2014-07-07T20:02:47Z2014-07-07T20:02:47ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.net6bae80bb0ee5aa590be3343ef2365b8e0b55bf19<div class="markdown_content"><p>Here are some cases where Maxima is not orthogonal and could be improved:</p>
<ul>
<li>Preserving bfloats in rational expressions (keepbfloat?)</li>
<li>Preserving floats and bfloats in matrix expressions</li>
</ul></div>Proposed work areas modified by Stavros Macrakis2014-07-07T20:01:06Z2014-07-07T20:01:06ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.net44e8af753beaf39e62e7e6020ac17889a6ed7a5d<div class="markdown_content"><pre>--- v1
+++ v2
@@ -1 +1,20 @@
Here are some areas that look like they're worth working on. For each area, there is a one-line summary, with a link to a detail page. Each detail page should be treated as a Wiki page -- not a blow-by-blow record of a discussion, but a summary of consensus. Discussion itself can continue to happen on the maxima-discuss mailing list.
+
+* [Orthogonality] Correctness and orthogonality in simple cases.
+
+* [Documentation] Better documentation, especially for new users. Here's a proposal to kick-start this: we should document every answer to a user question on the mailing list on the wiki, and develop over time a logical organization to the wiki.
+
+* [Solve] Coverage of more cases in solve, perhaps starting by using Barton's %solve.
+
+* [Comparison] Coverage of more simple cases in comparison (e.g., x^2+2*x+2>0 or sin(x)>=-1).
+
+* [Limits] Improve limit, probably incorporating tlimit into the main algorithm.
+
+* [Formatting expressions] Review the various existing ways of reorganizing expressions -- in particular, contrib/format should probably be upgraded to share after we review its functionality, test suite, etc. Beyond that, can we unify semantically or at least document the various ways of manipulating expressions?
+
+* [Pattern matching] A more general (though probably slower) and easier-to-use pattern-matcher (with backtracking, unification, etc.).
+
+* [Non-commutative operators] More functionality around non-commutative operators, both in the general case and for specific cases like matrices. Generalize the non-commutative system so that there can be more than one such operator -- each with different properties.
+
+* [Symbol and function properties] More general and more consistent handling of properties.
+
</pre>
</div>Proposed work areas modified by Stavros Macrakis2014-07-07T19:58:13Z2014-07-07T19:58:13ZStavros Macrakishttp://sourceforge.net/u/macrakis/http://sourceforge.net360a43981a5454382c8ea258ba3f3fc7022f8161<div class="markdown_content"><p>Here are some areas that look like they're worth working on. For each area, there is a one-line summary, with a link to a detail page. Each detail page should be treated as a Wiki page -- not a blow-by-blow record of a discussion, but a summary of consensus. Discussion itself can continue to happen on the maxima-discuss mailing list.</p></div>