On Mon, 29 Sep 2008, Riccardo (Jack) Lucchetti wrote: This seems very useful to me: it's now added in CVS. Allin.

Thanks very much for this. It looks very good. Quite often, I see in the default regression output that the names of the stats and the values below them are not aligned correctly. I think the spaces between the titles are predetermined and fixed in the pot file, no? So, I was wondering, can this be modified and used for tabulating all types of regression results by default? This would increase the overall polish of the program by a small but signifcant margin. Cheers Talha On Mon, Sep 29, 2008 at 2:39 PM, Riccardo (Jack) Lucchetti <r.lucchetti(a)univpm.it&gt; wrote: -- "Remember not only to say the right thing in the right place, but far more difficult still, to leave unsaid the wrong thing at the tempting moment." - Benjamin Franklin (1706-1790) --

I scanned various tables and realized that misalignments is not "quite often" as I remembered. Besides, some of them only affect the Turkish translation. Here is a small report: 1)-The "confidence intervals for coefficients" window from the analysis menu can use some formatting. The headers can be misaligned and the decimal dots in the upper and lower values columns are not vertically aligned. 2)-For the forecasts window, open Greene7_8, run an ols with G as the dependent and no independents (just intercept). Select forecasts and just click OK. The dashes between the lower and upper values are misaligned. 3)- In the normality of residual window, "interval midpt frequency rel. cum." look a little misaligned. 4)- In the Mahalanobis distances window, the chosen variables are listed vertically which gives the impression of a misaligned table when there is only one variable chosen. Could be better if the variables are listed horizontally as: "Mahalanobis distances from the centroid using the variables: X, Y, Z, ..." and the table has headers. 5)- The spectrum window can sometimes look a bit messy. For example, compare the sample periodograms (not log scaled) for CAN and D1 for the AWM dataset. 6)- The Turkish for range is "aralık" and mean is "ortalama." Because of the difference in the number of letters, the range-mean table looks misaligned in Turkish. (Also, please check the range-mean graph for abdata. When the value is too big (observation 14), it causes a small misalignment.) 7)- It is similar for the "display actual, fitted, residual" window. In Turkish, fitted is "yakistirilan" which is considerably longer and creates a problem in this table. 8)- The misalignment in cross correlogram was my fault in the Turkish translation, which is fixed now. 9)- In the "display values" and the forecasts windows, "obs" (the title for the observations column) is hard coded and needs to be translatable. 10)- A bug: Open abdata, run an ols regression with IND as the dependent variable. Choose "influential observations," and the program crashes: *** glibc detected *** /opt/gretl/bin/gretl_x11: free(): invalid pointer: 0x0a0877b8 *** On Tue, Sep 30, 2008 at 6:07 PM, Allin Cottrell <cottrell(a)wfu.edu&gt; wrote: -- "Remember not only to say the right thing in the right place, but far more difficult still, to leave unsaid the wrong thing at the tempting moment." - Benjamin Franklin (1706-1790) --