I have also collected a few suggestions:- Project autosave option in some minutes interval, to protect against data loss when program hangs.
- Always-on-top option for windows.
- New labtalk commands for project explorer organization.
- More labtalk commands to deal with data grouping properties in a layer. Now the only possibility is to group or ungroup *all* the plotted...
- From-to plot specification for function graphs (they are currently plotted over, and only over, the whole x-axis).
- Apparently uniform density of points for function graphs also in log-scales.
- In axis breaks: possibility of setting equal from-to break, with the purpose of changing only from linear to log scale in the same graph. Also change from log to linear should be enabled, currently it is not.
- Background invisible grid to fit objects to.
- More lines at specific, programable positions, not only x=0 and y=0 (y=constant is possible with a function and x=constant with a drop-down-line scatter plot, but complicated).
- New "yindex" (yindex1) function, similar to "xindex" (xindex1), returning the index of the first position of a y-value equal or smaller (larger) than a given one.
- More easy and effective operations on selections (instead of on whole dataset): now I need to write all the following to make things work, (and in other cases, they don't work at all!)
worksheet -s 2 0 2 0; // select 2nd column, named T
selc1=2; // again select 2nd column!
selc2=2;
%c=%h_T; // four different instructions to make an effective selection for multiple regression!!!
run.section(wksStat,MultiReg);
- Why not longer (and then, more) names for string variables? Why can I have thousands of numeric variables, but only a small bunch of textual ones (that are always rewritten without my control by the system or other macros)? Memory should not be a problem, as the Windows itself needs huge resources...
- Possibility of having a string as a substring notation argument (see http://www.originlab.com/bulletin_board/Forum1/HTML/000598.html)
- Creation of a chr() function, inverse of the asc() function.
- Personalized fonts on script, notes and result windows.
- Larger "Label Control" windows for more comfortable script editing (otherwise I have to increase the size by hand every time!) A "remember size-position" option would be OK.
- Non-printable object option.
- object.dx and object.dy properties, for indicating object width and height, produce bad results when using scales different from linear (e.g. log). Other syntax, indicating left, right, top and bottom coordinates, would give better results, as with layers.
- Combination of just some events for running objects script, not only "All".
- New "Project close" event. (Can "Window create" be used as "Project open"?)
- Specifying number of ticks in log axes is difficult now. "Increment" parameter is useless in this case (try to set it to 0.1 !), and "# Minor ticks" gives strange results. This should be corrected.
- About the analysis: there are good algorithms for uniformly spaced data, but there should be also the equivalents for randomly spaced data (e.g. smoothing, filters and calculus).
Maybe more another day.
Thanks for receiving the feedback,
jose.
P.S. I've been editing this message to add more items. For future modifications I'll write new posts, it would be easier to see the new ones.
[This message has been edited by Jose (edited 07-18-2000).]
