Tabbed objects vs non tabbed?

[DaveM]

For fun,
i took my biggest form and converted it and the supported form to tabs. It works nicely, but does anybody know why it takes several seconds to open rather than a split second?

I have included several pictures(4) and can you tell me which is tabbed and which is not?

cpu usage: tabbed 51% to open, nontabbed 26%to open

dave M

[Doug Page]

Re: Tabbed objects vs non tabbed?

My guess:

1 & 2 are not

3&4 are

As far as your question regarding the speed of opening the page, there has been quite a bit of discussion on the board as to the extra overhead that tabbed objects require. From what I remember, the more objects you put on the tabbed object the longer thngs take to operate.

[G Gabriel]

Re: Tabbed objects vs non tabbed?

As Doug said, 1&2 non-tabbed 3&4 tabbed.

[Cheryl Lemire]

Re: Tabbed objects vs non tabbed?

Hi Dave,

A tabbed object is an additional object on your form that needs to load. There is probably a bit of code behind that object that designs the multitude of tabs that you are using and locations of all the objects. To me, this would make load time a bit slower as Alpha has a lot more to decipher.

[DaveM]

Re: Tabbed objects vs non tabbed?

So, Why use tabs?? Some swear to it and some hate them.

After testing on a VERY fast standalone computer, it took 7 seconds to load the tabbed(average) form and less than a second for non tabbed form on a first start. After the first start, the tabbed form took a little less time(about 4-5 seconds) while the tabbed form didn't change noticeably. This wa using 2 instances of alpha running concurrently and 2 different folders.

Switching from tab to tab is a lot faster than opening and closing forms.

Doug & G were correct. It does show how you can open a form to look very similar to tabs.

Dave Mason

[MikeC]

Re: Tabbed objects vs non tabbed?

Dave,

I use tabbed forms quite a bit (some for just quick sorting) and they load virtually in about the same amount of time as my non-tabbed forms.

The only difference I can see is I usually only have 3-5 tabs and do not have much data...a few hundred records would be the most (maybe a few thousand after many years).

I have noticed even on my simple forms that when I have more than a few color calcs, especially in a browse, that the time loading and the screen "flickers" increase.


Curious--
Do you also have additional calculations on the tabbed forms (including color calcs) ?

[DaveM]

Re: Tabbed objects vs non tabbed?

Mike,

Good Morning.

I have many forms started from that main nontabed form. several of those have calculated objects that had to be brought into the tabbed form, There are very few color calcs. The fields are numerous on some tabs. the only calc added was to save between tabs. A lot of buttons were dropped when moving to the tabs.

The forrms are bassed on a set with about 10 tables(no grand children), and very little ref integrity. There are about 4200 records in the main table and som 20 times that in a few of the 1-many children.

Dave Mason

[MikeC]

Re: Tabbed objects vs non tabbed?

Thanks for the info Dave.

I guess my main point was just that....to compare the tabbed form load time to a non-tabbed look-a-like one I think the calcs would have to be removed in order to see what part the tabs themselves are playing in the time it takes.

I had a form with just color calcs in each of the browse columns and removing them cut the time in half for loading the form. But other factors could have played a part as well as I had removed some script that was tied into the calcs...but this only seemed to make the screen flicker more than before and not affect the load time.

Always nice to find out these little tidbits about how things are done and what affects what and such.

[Cheryl Lemire]

Re: Tabbed objects vs non tabbed?

I happen to one of those that loves the tabbed objects on forms and use them frequently when needing to display a lot of data. I have on form with a tabbed object on it that has more than 160+ objects.

I have another form that has a 10-tab tabbed object on it with just over 100 objects. The form is based on a set with 9 children, 15 grandchildren, and 1 great grandchild. There are more than 15,000 records in just one table of the set and approx 80,000 records in all the tables combined.

There are approx 40 fields, 8 memo fields, 9 embedded browses, 8 embedded browses of memo fields, 37 buttons (with 3 events tied to each of them) and a sub-tabbed object on the main tabbed object with 5 tabs.

This form takes less than 3 seconds to load and this is on a system with only 512MB of RAM and 1500MHz processor.

Although the tabbed object itself may take slightly more to load, I think the types of objects on the form or tabbed object play a much larger role in the load time. The calculated fields, as mentioned previously, are a major factor in the load time.

I agree that creating a form that 'appears' to be a tabbed object is possible. But the 'tabbed object' allows me an 'easier' way to achieve the same result and gives me a lot more xbasic options for control.

All I am saying is that I would not knock the tabbed object. The tabbed object, by itself, I doubt is the cause of the load time.

[DaveM]

Re: Tabbed objects vs non tabbed?

Cheryl,

Glad you showed up. I figured this thread would get your attention. Kinda knew you were doing tabs from earlier posts.

Was hoping to find out if you and anybody else had slow load times and what caused it so I would know where to start.

It is like the load time of all the different forms added together. I have 5 browses acrss all the tabs, main db has 115 fields, there are a lot(maybe 35) calculated fields and some are very complex(these are recalced on recor change which is very fast). There is nothing wierd in the startup. It does have a couple xbasic queries to do(but so does the non tabbed). There are 18 variables declared. The fields(not including the browses) approaches 420. Buttons are about 30 and most will do 3 - 8 things at a time. There are about 25 udfs and another 65 scripts run from various buttons or ondepart, onsave,etc from within the code. The udfs and scripts should not impact load time though.

Since this is first and formost a CRM for automotive. The first tab is devoted to a search browse. the second gets rudimentary info on buyer and cobuyer, 3rd gets credit info for both with buttons for the long credit apps some have to use( I would make this diff if I used the tabs), then we get the sold vehicle, next is the trades(any number), next money(this is a huge area with lots of calcs and posts), rest are pretty static.

I really am thinking of going to a tab, but if load time is going to be like this??? My customers will happily live with a 1/2 second load time between forms, but will balk at a 5-7 second load time to get started even though the time could be the same overall.

Dave

[Cheryl Lemire]

Re: Tabbed objects vs non tabbed?

Hi Dave,

I can't pretend to know exactly what Alpha is looking for or processing at the time a form loads, so any comments I make are purely 'opinionated'. I doubt any of the UDFs or scripts you have in the actual button events have anything to do with load time, they would have impact during the actual event when it they are run. I imagine the events themselves need to be processed for content, but not the actions of the content. That is one reason I did not even begin counting how many UDFs and scripts that I have behind my 37 buttons, just the fact that they each had 3 events tied to them.

I would think that things that would effect the load time would be anything that has to actually be processed in order for the form to be viewed as intended. Opening of the tables involved with the links provided, placing the objects in the various locations, all calculated form objects would need to be processed, any UDFs/scripts in the forms OnActivate and OnInit events, any variables that have NOT been initialized at startup, any formatting calculations (colors, fonts, dimensions, etc). There are probably a lot of other things as well that I am not mentioning here.

One thing I would mention, my forms that have tabbed objects on them with a lot of objects and tables and records, are NOT used for data entry. They are strictly for 'viewing' of the data giving the user one place to go to get all the data they need related to a particular parent. With the exception of the search buttons, most of the buttons on my forms are for the purpose of loading 'individual' and separate add/edit forms.

The one form I mentioned with more than 160+ objects on it, takes 3-4 seconds to load. That particular form is actually a data entry form (I know I said earlier I do not use them for data entry, but there is always the exception to the rule). However, the only tables attached to the parent in this set are to 'display' only any phone data tied to the parents (these are read only browses).

I can tell you the reason this form takes 3-4 seconds to load has nothing to do with the fact that I have used a tabbed object or that there is more than 160+ objects on the form. The load time on this form is due to the UDF that I run which starts the add new record process. The UDF opens tables and populates variables and form field objects. This UDF has more than 700 lines of code in it. I have not tried this, but I would guess that if I removed this UDF, the load time of the same form would be a split second.

This form is used for data entry of an initial order which is written to a temporary table. When the user saves their changes, I then have another UDF that takes their entered data and creates new records in all the various tables that the data actually belongs in. This process also takes several seconds (it is also 700+ lines of code to process).

When users need to edit an order, they would load my 'view' only master form, navigate to or search for the record they are looking for, then selecting one of the various buttons to edit the data depending on what they need to edit.

Although doing things this way does not save on load time of the form (in your case it may if you are loading your form entering a new record and populating certain data up front) it does save time in many other ways.

If I have 3 people entering new records and 6 more users simply searching and viewing data, they are doing this in different sets/forms. Since my new order set/form is opening a temp table, the other tables that I open to retrieve data from are only open for a split second and not actually tied to the set. If I were to use the master form for data entry, I would have 9 people having all of those tables opened and being queried at the same time vs the 3 that I have now.

So now I have two major forms with similar tabs and objects on them, based on two completely different sets. This seems like twice the work, but it saves in processing time and network performance issues. It also helps maintain data integrity and minimizes user error. If I had these two forms designed for two different purposes (1 for viewing and 1 for data entry) combined with both purposes in one form ... then take the 3-4 seconds that each takes to load, the newly combined form would take 7-8 seconds to load.

If the individual forms load time really is an issue, one possibility that you could try would be to load the form at startup and hide it. This of course would add that load time to your startup time, but perhaps the users would except that? It is all a personal preference thing.

I honestly believe that if you pull out the 'data entry' aspects of your form and restrict it to viewing and searching, your load time would probably cut in half.

[DaveM]

Re: Tabbed objects vs non tabbed?

Cheryl,

I believe the complex calculated fields and short beginning query( selects user records and chooses open clients) may be suspect to load time. The calculated fields are necessary even for viewing and so is the short query.

The calcs do all the numbers required for a deal(all taxes, deal math with payments, profit and loss), as well a names, etc, so even a manager would need to see the results.

The query(in part) makes it impossible for one sales person to see another sales person's work. This is very important in this business.

I guess I will stay non tabbed for now.

Thank you very much. You really confirmed my fears. Tabbed would be nice, but I couldn't remember why I dropped it so uickly before.

Dave M

[Stan Mathews]

Re: Tabbed objects vs non tabbed?

I seem to remember a post from some time ago from someone who liked the functionality of tabbed objects but had some issues with them. He obtained the functionality of a conditional object with strategically placed buttons resembling tabs along the top edge. The buttons set values to a variable which drove the layer change of the conditional.

This may not be any better than what you have tried but is an alternative.


Link to discussion here.

[DaveM]

Re: Tabbed objects vs non tabbed?

Stan,

I have conditional objects in the tabs now. They were a little harder to set up than on a standard form, but work the same.

The tabs in my pictures (1&2) emulate tabs but are buttons that just open forms that are placed so they look like tabs(I could have done a better job). This works real well. the difference is, they have to be closed each time instead of clicking the next tab. I have not gone so far as coding an exit/quit form when clicking another button and not sure if I can. Not sure there is a need or desire either.

This morning when I started the computer, I started the tabbed form on a FIRST run and it took 12 seconds to load the tabbed form. After the first time it gets a bit better at around 10 seconds. This is on a stand alone test app(not shadowed).

If the speed were the same, I would do tabs. They are definitely harder to set up though.

Dave

[Stan Mathews]

Re: Tabbed objects vs non tabbed?

The discussion is not about conditional objects on tabs of a tabbed object but rather a conditional object instead of a tabbed object.