Poll: EG for programmers or not?

[ChrisNZ]

>SAS/AF happens to be why we don't use EG in my office

Yes, I am with you here, and SAS/AF is sometimes the only way to write a process.
Ever tried to write a recursive routine? Macros and call execute cannot recurse very far, only SCL can.

Solid recursion would be another welcome improvement when a revamped developer toolbox comes, or array functions similar to the SCL list functions or PHP array functions (though all this is not interface-dependant).


>Your "tips" on line numbers, obtaining path names, and getting variable names via copy/paste are spot on.

There are always ways to do things. proc contents or pathname() would work too here. Is that an improvement on the right-click contextual menu to display columns properties though?
More time spent faffing around running processes, going to the result window, deleting the tabs created, just to do trivial things.

How about improvements instead, like select a few names and formats, and drag and drop them from the columns properties window to the editor? Or the "Find column name" field finds as you type rather than requiring you to know the exact entire variable name? Or that the dialog can be resized so we can see names and labels at the same time?

By this logic, I don't really *need* the viewtable window (and its where; where also; where undo; show; hide; form view; change format; options) after all, I could just use proc print to browse data, right?

Where are the increased productivity and ease-of-use in that?

I feel a sense of people having resigned themselves to (my words) and having gotten used to (yours) EG. 'Or else' in some cases as Franco puts it.

Not quite the embracement or endorsement one would like to see, even if for some positive aspects have been mentionned. Not quite the raft of new features that have been missing and requested for years either.

Different features and fewer features instead. SAS has now been pleading for years for developers to use EG, and the migration is slowly happening. They wouldn't have to plead if they provided the right tool set. The move to EG reminds me of Rumsfeld's reply "As you know, you go to war with the army you have, not the army you might want or wish to have."

As far as I can see, EG doesn't answer my current needs:
-help me better manage the company's macro libraries, format libraries, code libraries;
-help me debug macros, help me find the important parts of my log,
-help me write code as I type, help me find an earlier versions of a piece of code,
-help me track code changes.
EG does seem to bring some positives to some of you, but it also creates more hurdles, and I don't need more hurdles.

I don't know about you, but my code is always wrapped in macros. Macros make maintenance so easy.
Even simple things like titles use macros, as in
[pre]title2 h=%fontsize(title,&FontSizePt) [/pre]
with
[pre]%macro fontsize(usage, pt);
.....
%let __dev=%upcase(%substr(%sysfunc(getoption(device)),1,3));
%if &pt=12 %then %do;
%if &__dev=ACT %then %do;
%if &usage=TITLE %then 10pt;
%if &usage=ANNO %then 1.5;
%if &usage=AXIS %then 10pt;
%end;
%if &__dev=GIF or &__dev=PNG %then %do;
%if &usage=TITLE %then 12pt;
%if &usage=ANNO %then 1.5;
%if &usage=AXIS %then 12pt;
%end;
....[/pre]

because pt font size is not available to annotations (except via goptions), and different devices interpret the same size differently anyway. Now I can change the font size or device across all my outputs from one location, without having to worry about annotations or the chosen device. Likewise ODS destinations have their quirks. EG doesn't help me manage that kind of time-saving flexibility, does it?

I am very ready to change, don't get me wrong. I want to drop the old DMS. I want modern features like a Find dialog which finds and highlights as you type. I am ready for server-side processing with just an interface on the PC if that's the way things will go. But I don't want change for the sake of change, I want change for the (much, if possible) better.

>Many of us walked away from this year's SAS Global Forum, in Seattle, WA, with the same feeling of abandonment.

I too wonder whether sas might be overlooking its developer base. How wrong would that be. EG end-users can be retrained in a day to use another (not as powerful) reporting tool if a company decides to drop sas. It is the sas developers and the complex, flexible and fast processes they create that make sas a company's valued asset, not the reporting tool, and helping the developers should be a top priority.

Alright, enough arm waving, you got my point, so I'll stop my tirade/plea here. I hope it was somewhat constructive, in pushing my ideas for (what I see as) a better future.

I'd still be curious to know the uptake of EG among Sas Institute developers: they too use lots of macros, and they also want speedy code.

[nrose]

Hi again,

You can still use macros in SAS EG, and I often use EG to do the heavy lifting so as to produce code templates that I can then modify into macros.

Also, I think that if possible, people should seriously think of migrating away from AF, and into .NET development. It is becoming easier for people to extend their skillset into .NET with the code templates and Visual Studio Express being such an easy development environment. Once you move into .NET, you can do just about everything, and SAS EG integrates nicely.

I am one that recommends that SAS focus their efforts into SAS EG, and .NET integration, and away from legacy products that have long passed their used by date.

Nick

[ChrisNZ]

>I am one that recommends that SAS focus their efforts into SAS EG, and .NET integration, and away from legacy products

I am with you overall except that I reckon a new product SAS Enterprise IDE is needed beside SAS Enterprise Guide.

Also, I forgot to thank the forum admins for tolerating what could be seen as low-level dissent by people not open to suggestions and discussions. 🐵

[art297]

Chris,

You can't be in total agreement with both Nick's and my responses .. we are, in some ways, diametrically opposed. Many of us have spent numerous hours and financial resources on developing what we currently have.

No I am not a total dinosaur and fully realize that better offerings might be accomplished with newer technology and, in fact, would welcome such advances.

However, one should never throw a baby out with the bath water, leave their older clientle out to dry, or stop improving what they have at the expense of developing something that MIGHT be better.

Yes, SAS could spend less on things like SAS/AF improvements in order to spend more on newer technological advances, but NOT simply stop trying to improve the products that they sold us on and are currently using.

I REALLY don't think it is in SAS's best interest to have many frustrated clientle.

p.s. I, too, am impressed that they haven't cancelled (or too my knowledge edited) this thread. Otherwise I'd jump on the soapbox and argue for a movement toward only using user controlled forums.

[DrAnnmaria]

While I do agree with some of the drawbacks you see, for example, the log window is annoying, I find EG very useful. I often use it like this:

1. Do all of the data management _ I think_ I need using SAS 9.2, merge datasets, recode data, blah blah blah.

2. Do graphs, tables, and common statistics, e.g., ANOVA in Enterprise Guide or turn it over to the client and let him/her do it to their heart's delight.

3. For the one or two things I always forgot - changing all character variables to upper case or whatever, I open a code window and write a few statements.

4. If I have a complicated program, I may do most of it as SAS program and just open the code window, use a %include statement to read in the vast bulk of it that I already tested, add a few statements for modification on the fly.

[GoodPeace]

Chris@NZ, if I ever meet you in person, I'll buy you a beer for your excellent postings in this thread. Well behaved, not at all like a rant. It's all very important issues for the SAS user community but even more so for the SAS Institute people (managers as well as software developers). Kudos to those admins for not shutting this down - SAS seems to be moving towards that fancy social web thing all the kids talk about... 😉

Now, back on topic:
I would love to love EG! I eagerly await my chance to get the upcoming features in 4.3. Code completion/predictive typing - YAY! Being able to get that process diagram overview of programs, especially colleagues' handed-over programs - awesome. That's what I do in my head for small programs and for larger programs by printing out several A3 size pages of pt 7 sas code and drawing lines and boxes in hand all over them. Having wizards and tasks for those procedures I can't remember the details of - great for my aging memory.

I really do look forward to it, and I really appreciate and understand the great benefits of the kind of coding/developing EG encourages.

So all others, don't bother painting all those up for me - I know about them and acknowledge them already. We agree, it's good.

However, I have used DMS for quite a few years now, and there seem to be all these small things that I can't do, whenever I try to use EG (4.1 up until now). These things are rarely big bugs, but they're still terribly annoying and counter-productive. Linux communities use the phrase "paper cuts" for all those tiny missing/buggy features that needs fixing, but never get the attention of the software devs. I don't know if these things are like paper cuts, but some of them seems so obvious and easy to fix, that I' amazed they're missing.

The dataviewer including command line and quick where selection (plus some others mentioned earlier) is a must-have. When investigating various data sources, figuring out how to put things together, looking up that one customer which appears a certain way on the front-end screen of the production system, etc etc, I can do it in a snap in DMS. I have some basic keyboard shortcuts to get from one table to my explorer (two-paned folders|content btw), open a new dataset, browse to the right, go to the command window, use where to view on certain row, flip to the other open dataview, do another where command, etc etc. I almost never touch the mouse meanwhile, and that is the main reason I can do it so quickly. Maybe some Counter Strike pro player could use the mouse fast, but a dull office chimp like me, do stuff fast with keyboard, not mouse.

Doing the same thing in EG is slow and painful and the paper cuts my fingers to blood. Why not make it easy, SAS?

If I prefer to use F8 or ctrl+L or shift+alt+B or whatever to get straight from my code editor to a two-paned explorer window with perfect TAB/arrow shortcut functionality, why can't you make that work? This is simple, it's mostly in the interface. It's not rocket science - almost all other applications in various fields across Windows and Linux desktops have had such features for ages. I suppose this is why Chris@NZ mentions Visual Studio IDE.

As others have said (in less harsh words), DMS is a lousy piece of outdated and sub-par interface not at all suitable to modern PC users. However, it does allow some good things for us, and we've come to like those things. If SAS wants us to go to EG, then why not make it worth the shift? Make it shine, make it flexible, make it user-friendly, make it customizable, make it modern.

Someone mentioned Excel - this reminds me of my old happy horse: SAS should go headhunting for some of the interface designers from Redmond, they seem to understand what a user-friendly interface is all about. The joke is getting old though, since every app dev and his younger brother are better than both SAS and MS nowadays 🙂

[art297]

One of us have to get in line for I, too, want to buy Chris a beer or two. In fact, my thanks to virtually everyone participating in this thread for being so civil throughout this discussion.

The one last point I want to make before opting out of the thread, though, is the point I originally wanted to make but don't know if I really ever got the message across .. particularly to SAS.

First, I really do hope that SAS enhances EG well beyond the point of not giving you paper cuts. Invincibility might be a bit much to hope for, but I wish for as close to that as they can provide.

However, I have to speak for those of us who will never get to enjoy the fruits of those efforts, but will still be using SAS for the next five years or so.

As I've mentioned, my office relys on an AF application and, to date, nothing has been offered that can replace it. What I haven't mentioned is that a large part of that application was built in the SAS6 days and SAS has never offered anyway of automatically upgrading SAS6 AF modules even though they drastically improved it in SAS8. Unfortunately, they didn't upgrade many of the various components they provided with it, thus the entire Toolbox project can only capitalize on SAS6 modules.

I'm sure my office is not the only one that is still living with such legacy code. And legacy code, fairly soon, will have to be replaced. My one wish is that SAS will realize that and make it easy for my predecesors to capitalize on the many years of work that went into developing what we currently have. Conversely, my ultimate replacements can spend ten to fifteen years just getting back to our current levels, whether that be with a future version of EG or some non-SAS language, but why should that have to be!

That is the final question I have to ask.

Art

[ChrisNZ]

Alright, I can't say no to a free drink can I? So airline agents, here I come! Where to, the tickets?
Seriously, thanks guys for the kind praise. I just hope something will come out of this discussion.
Paper cuts? I never heard the term before, but it seems right. Like the variable list in viewtable that is just 8-character wide and can't be resized (Yes, I did report this feature with tech support years ago), or when you hit Ctrl-H, fill in the Find field, then Tab, then fill in the Replace With field ... except Tab doesn't take you to the To field, and you are typing away into the vast empty cosmos. Sigh.

If sas kills the DMS in V10 and hopes everyone will be content to use EG or a close sibling, they are in for a surprise imho.

>If SAS wants us to go to EG, then why not make it worth the shift? Make it shine, make it flexible, make it user-friendly, make it customizable, make it modern.

Exactly, make something people will want and they'll embrace it and dump the DMS in droves, which they aren't exactly doing atm.

> The joke is getting old though, since every app dev and his younger brother are better than both SAS and MS nowadays.

That's a bit harsh, and sas has the extra constraint of making multi-platform interfaces/apps.

But you do have a point. How patronising of MS to force a change of interface (regardless of the ribbon's merit, and the jury is still very much out) onto unwilling users (at huge training costs) for no good reason.... Message was edited by: Chris@NewZealand

[GoodPeace]

The free drink invitation is open in Copenhagen, Denmark - however, that's quite a long journey for a free drink 🙂

Agreed, tab-functionality is not at all working as expected in EG, and that's an often seen paper cut, which can't be that hard to fix.

It's basic stuff like that, where I see lots of free open source software being developed towards both windows and linux environments with features and usability way beyond what SAS is doing - that's the reason for my, still joking-like, mention of every younger brother app dev out there.

Now, the EG 4.3 has landed at my site, I hope I can find some time to play around with it.

[pablo_joerger]

I am using SAS several years. In the 80's on DOS, these days on Windows.

EG is really an improvement to perform efficent and on time good and validated programming using SAS. The best at the time being is the integration you have these days with SAS 9.2 and EG 4.3.

I feel it is a great move forward and makes life of SAS programmers and users more and more comfortable and accesible.