Suggestions for Stata 15

[Phil Bromiley]

I know its early to start this, but maybe that is the right time.

Two suggestions:

1. In the do-file editor, it would be nice to be able to split the screen. I often find myself wanting to see exactly how I named or defined variables earlier in the file when I'm working later.

2. In the documentation, I often use the subject index. However, getting to it requires several steps. Maybe the subject index should be at the top level (like all the other books) so it is more easily accessed.

[Nils Enevoldsen]

Also refer to: Wish list for Stata 14.

[Richard Williams]

Yes, everything that didn't get done from the Stata 14 wish list automatically carries over to the Stata 15 wish list. ;-)

This doesn't really need to wait for Stata 15. But I would like as many manuals as possible for old versions of Stata kept available online. These can be helpful for determining how a command works when version control is used.

[Giulia Zilio]

and please create an outreg2 (word) command for MAC. Thanks

[Nils Enevoldsen]

Giulia: outreg2 is user-written, and StataCorp won't integrate it. However, -outreg2, word- already works on OS X. I don't understand the suggestion.

[wbuchanan]

More importantly, however, is that Stata has had natively integrated support for both MS Excel and MS Word since Stata 13. I would still like additional documentation for graphics to make it easier for the user community to help contribute to Stata graphing capabilities. It would also be nice to have a way to allow user defined functions/factor variables which could also potentially be covered in additional documentation. In terms of bigger asks, it'd be nice to see JDBC become integrated in Stata as well as alpha layer transparency and/or other options to allow layering graphs without being quite so verbose (e.g., adding an over option to tw connected/scatter that could allow users to map distinct colors and/or shapes to the values of a variable). Latent class analysis would be a nice addition as well.

[Andrew Lover]

An -eform- option for Bayesian models (as appropriate) would be fantastic.

[wbuchanan]

I'd also like to add expanding the Java API and the documentation/examples for the API/Stata functions involved in its use. An unrelated additional request would be allowing the st_data() and st_view() mata functions to accept both string and numeric data.

[Sergio Correia]

allowing the st_data() and st_view() mata functions to accept both string and numeric data.

I'm guessing they had performance reasons for splitting the string and numeric cases into st_sdata() and st_data(), so I'm not entirely sure if that's feasible (perhaps a wrapper can call one or another depending on the input?)

[Erika Kociolek]

See Wishlist for Stata 15.

[Gordon Ballingrud]

More convenient post-estimation commands for bayesian output, like marginal effects and predicted probabilities, would be perfect.

[Joseph Coveney]

Important: ability of bayesmh to work with instantiated Mata classes directly.

Nice-to-have: a bayesmh postestimation command to find the multivariate mode of the posterior.

[Johhn King]

Adopting some flavor of software licence server, and providing a true concurrent usage network licence option, would be highly advantageous for the irregular way in which we use Stata.
This would also provide the ability to forecast our licence usage and future requirements.

I would also like to see greater support/documentation/integration with newer hub/spoke deployment technologies like App-V and Citrix XenApp.

[peter cel]

I'd like a popup box facility in Stata15. so for example if you start typing "eg" a pop-up box suggests egen and so on. if there are more than one possibility arising from what you initially type then the pop up box offers those possibilities. Many software (e.g., Mathematica) do this, and I think it helps