Subscribe to this thread
Home - General / All posts - Creating Centroids for Labels
dchall8
1,008 post(s)
#18-Jan-18 16:22

In watching the video, Manifold 9 - New Labels Features in 9, from 0:40 to 0:51 the centroids drawing is described as taking about a second to generate using the transform tool. When I tried that the centroids replaced the original area leaving me with nothing to put a label on. After clicking around I never found an easy way to make centroids and add them to a Centroids drawing. So I did the following...

  1. Left-click to select the table with the areas I wanted to label
  2. Click Copy icon
  3. Click Paste Icon
  4. Right-click new table and create new drawing based on that table
  5. Open new drawing
  6. Zoom to area of interest
  7. Select specific areas to make centroids (point and ctrl-click)
  8. Select inverse
  9. Delete selection
  10. Click Contents Pane
  11. Click Transform (Geom selected)
  12. Template tab
  13. Select areas in drawing
  14. Click Center, Inner in Template
  15. Click Update Field

BOOM! Just 15 simple steps. Since my method took longer than a second, I have to ask what the shortcut is to make centroids and keep the original parcels intact?

Dimitri


7,413 post(s)
#18-Jan-18 17:12

Tables? Select Inverse? Delete Selection? You don't need any of that.

You are way over-complicating this. It is far easier and faster than your list.

Suppose you have a map of areas. You want to create centroids. Easy:

1. Open up the map full of areas.

2. In the Transform panel, choose Center, Inner. (time budget: 2 seconds, including time for a quick click on the Transform panel if you don't have it open.)

3. Press the action button set to Add component. Done. (time budget: 2 seconds)

I just did it, and from staring at the map to having centroids created took me less than five seconds. The actual generation of centroids took one click, less than a second, with the rest of the time being a click on the contents pane to bring it to the front, a click on the transform panel, a click on the Center, Inner template to choose that. Unless your finger is broken three clicks takes maybe a second or two. No change to the original parcels, since you create a new component for the centroids.

If you repeat the example in the video, the centroids drawing really is that quick to create. If you're really slow as a beginner, well, maybe it would take ten or fifteen seconds to manage the very few clicks involved but it doesn't take much learning to get that down to under one click per second for really simple things like this.

By the way, I guess from your list that you are not out to create centroids like the video remarks but to do something different, like making centroids for a selected subset of areas. No problem as that is fast and easy to do also.

Take a few seconds (all it takes) to ctrl-click the areas for which you want to create centroids, in the Transform pane check the Restrict to selection box, and then it is the same one click to create Inner centers, or centers of weight, or a regular center. Just one click to check that Restrict to selection box which is no big deal, one second max.

Just for the heck of it, here is the full click roster for the original task:

click - Bring Contents pane to the fore.

click - Bring Transform panel to the fore.

click - Choose Center, Inner template

click, click - Change action button from Update Field to Add Component.

click - make the centroids (done).

That's six clicks, the first three of which are a cluster of related moves that a person who knows 9 can do in one second, and the second three of which are also a cluster of related moves that also can be done in a second. A fast operator could do this in two seconds, not the five a slow poke like me takes. :-)

If you want to choose some subset of areas, well, that's extra on the time budget but that would be true of any technique: if you are staring at a map thinking that you want some centroids for some areas but not other areas, well, your thought process is part of your time budget regardless of what tool you are using.

What I will say with 9 is that whatever thought process you are using to decide what areas you want to pick out, a tool that makes it fast and easy to apply a wide variety of selection methods in super fast ways is going to help that process happen quicker.

Mike Pelletier

2,122 post(s)
#18-Jan-18 17:51

The Update Field / Add Component button is a crucial tool to remember in 9 and it kind of is hidden in the button that runs the transform. Perhaps set all the functions up so it's more clear the difference between the option buttons and the button that runs the transform.

Also could the options be enhanced to allow transferring the result to another component (Mf8 made that easy) or add/subtract the result to the current drawing? Maybe set it up with a target button and an operation (add/subtract/replace) button.

KlausDE

6,410 post(s)
#18-Jan-18 18:19

For me the standard position for the action buttom is the lower right corner, only a cancel wound be even more to the right. Options I would expect above or left of the action button. As an example see the bottom line of the file dialog.


Do you really want to ruin economy only to save the planet?

tjhb
10,094 post(s)
#18-Jan-18 23:30

I wasn't sure about this composite control at first either. Adam confirmed that it is a standard Microsoft Win32 control--but unfamiliar because not widely used.

Personally I think the issue is only one of familiarity. It needs to be drawn to the attention of new users, obviously, but after you get used to it, it starts to seem very clean and efficient. Anyway that has been my experience. I now like it.

Dimitri


7,413 post(s)
#19-Jan-18 02:25

I, too, did not like it at first. I would have preferred an option box with a pull down list of actions such as Update Field, Add Component and so on and then next to that an Apply or OK button. But... doing that would have had two downsides compared to the current control.

The first is that using two controls uses expensive real estate for two controls to duplicate what one control does just fine. The second downside is that it becomes easier to click OK or Apply when what is in the option box is incorrect. That is slightly harder to do when the button you press is obviously captioned either Update Field or Add Component.

Given the Transform panel is a "live" control where a command can reach out into billions of records in a database that may not allow changes, and where very many people (imagine altering a database of prices for flights in an airline company database...) might in milliseconds be using those changes, well, there's a non-negligible benefit to the latter interest in having the button that does a command be captioned with what it does.

Anyway, now that I've gotten used to it I like it, and, oddly enough, even though I work with three monitors and thus am not usually cramped for screen real estate, I like it that it saves real estate as well.

Mike Pelletier

2,122 post(s)
#19-Jan-18 02:32

Well I suppose you and Tim are right, but a super clean interface is a double edged sword. It is very nice when you know what your doing, but it can be maddening when you don't. It feels like a room with no door to get out and just a thick manual laying on the floor :-)

adamw


10,447 post(s)
#23-Jan-18 13:12

Also could the options be enhanced to allow transferring the result to another component (Mf8 made that easy) or add/subtract the result to the current drawing? Maybe set it up with a target button and an operation (add/subtract/replace) button.

You mean something like -- select a couple of areas, then create centroids for the selection, *and* put the resulting points into the same drawing rather than into a new drawing or replacing the areas? This is doable, yes. Not for all transforms, but at least for those that support Update Field.

Putting the result into a specific existing component is also doable, but we'd rather make it easier to use copy and paste first.

dchall8
1,008 post(s)
#18-Jan-18 18:46

I tried using the Add Component before writing the first time. Nothing happened. I'm attaching pictures to show the before and after results. So I tried it again very carefully following your 3-click process. Centroids are clear in the Before image but they disappear in the After image. When I deselect the parcels there are no centroids. When I check the table to see if there are any new records, there aren't. When that approach failed I got desperate and went to the brute force method posted above. I'm using Build 9.0.164.3 on Win 7.

Ultimately I would create centroids and labels for all 32,000 parcels all at one time. This was just to practice making labels. One thing that is an issue at the current state of development is that I want labels to turn on and off when I ctrl-click on individual parcels (or otherwise select them). In M8 I theme the font color based on the Selection (I) field for the parcels. The way labels work in M9 they are not tied to the parcel object but to the centroid object. When I select a parcel I would not be selecting the centroid, so if there was a Selection (I) equiv in M9, then clicking on the parcel would not activate the theme in the Centroids layer. In fact I would normally have those centroids turned off.

Rhetorical question: In the Transform panel, why is there a dropdown menu for a binary selection (Update Field and Add Component)? A dropdown requires 2 clicks to change the value plus the extra click to activate the button. If it were a radio type button, or two buttons, changes would require only 1 click.

Attachments:
After clicking Add Component.jpg
Before clicking Add Component.jpg

TAL7 post(s)
#18-Jan-18 19:34

A new drawing component and table with centroids will have been created in the Project Pane. Add the new drawing component to the map.

dchall8
1,008 post(s)
#18-Jan-18 20:43

Thank you, TAL. I have two new tables and drawings, so I have a record of how many times I tried this. I knew the 15-step process was not what they were referring to in the video.

Dimitri


7,413 post(s)
#19-Jan-18 09:17

so I have a record of how many times I tried this

"tried this" ? Tried what? Repeatedly trying grossly wrong procedures is not efficient. What is efficient is to try a process that you think should work once, keeping careful notes.

You can then get on the forum and write a post that says:

1. Here is my data (schema and all).

2. This is precisely what I want to do.

3. Here is the step by step procedure I did (leaving nothing out).

4. At this step... here is what happened (in detail) as compared to what I expected (also details).

.... then people can save you from repeating wrong procedures and offer a step by step process to accomplish what you want.

Dimitri


7,413 post(s)
#19-Jan-18 09:10

I tried using the Add Component before writing the first time. Nothing happened.

Sigh... the above tells us nothing. "I tried using the Add Component" does not tell us what you did. "Nothing happened" also is imprecise. I get the feeling you haven't read up on how the transform panel works, and what the add component button does, that perhaps you expect it to add a layer to the map. Not what it does. It creates a separate component in the project. It doesn't modify a map you are working with.

Ultimately I would create centroids and labels for all 32,000 parcels all at one time.

Well, no, you actually want to do something more than that. Creating centroids and labels for all 32,000 parcels takes seconds. Follow the five-second process I gave to create centroids. Right click on the centroids drawing thus created and create labels (takes another 2 or 3 seconds). Done.

I want labels to turn on and off when I ctrl-click on individual parcels (or otherwise select them).

Ah, well, that's different. Also possible to arrange but you would structure the project differently in that case, for example, two geom fields in the same record, one for areas and the other for centroids, so that both are selected whenever the record is selected. Easy to do.

This is, by the way, why consultants get a specific description of work to be done and then they charge an arm and a leg for change orders. When the client says "I want to create 32,000 centroids and then labels on those centroids," the consultant says sure, has the client sign, and then the consultant delivers the job.

When the client next says, "oh, no, not exactly, I want to create labels for selected areas..." the consultant can say, OK. Sign this change order and I'll do that for you too. Adds a quick use of the Restrict to selection box.

And then when the client says, "oh, no, that's not it either... I want labels to turn on and off when I ctrl-click on individual parcels or otherwise select them..." well, at that point the consultant is very happy because he gets paid for yet another change order.

A good consultant would sit down with you to understand what you are trying to do from first principles, what the business objective is, so that instead of decoding half-steps taken on the basis of a different product they could just set something up for you right away in the product you are using.

A greedier consultant might just take each comment you make, charge you for that implementation, and then when the "oh, well I wanted something a bit different..." would just charge you a change order.

By the way, before commenting on what you think is or is not possible at the current state of development it would help you avoid false starts if you first learned how to use what the product can do at the current state of development. The moment anybody starts thinking "oh, that's a bug" or "oh, it can't do that" when the problem really is a lack of know-how, the learning process stops.

The learning process is also important because it could be the way somebody approaches a project is based on limitations which are not there. I don't know what the specifics are of what you want to do, but any manual turning of labels on or off for 32,000 parcels gives me the vibe of a missed opportunity to use automation instead of manual effort. There is just something about poking at individual objects in a sea of 32,000 that smacks of inefficiency.

The most efficient, easiest, fastest workflow usually comes from using a better algorithm. The best algorithms tend to come from well-informed and expert users. Considering a task with full awareness of what a tool can do will often show ways of proceeding that can save enormous amounts of time and effort.

tjhb
10,094 post(s)
#19-Jan-18 09:57

... you would structure the project differently in that case, for example, two geom fields in the same record, one for areas and the other for centroids, so that both are selected whenever the record is selected. Easy to do.

Very nice Dimitri. Elegant!

dchall8
1,008 post(s)
#19-Jan-18 18:22

Sure! Why didn't I think of using two Geoms? Because I have an extremely limited idea of what a geom is. It seems to be an amorphous container for all the secret graphical intrinsics I can no longer see. There are tons of references to geom in the voluminous manual, but a quick survey of them did not reveal a helpful definition. Nor do I think I really need one for the level of GIS I do. At least I haven't needed more knowledge in that area in 13 years of using Manifold.

I tried using the Add Component before writing the first time. Nothing happened.

Sigh... the above tells us nothing. "I tried using the Add Component" does not tell us what you did.

I get no credit for the 15 step process I innumerated, but I catch crap for not going into the details of pushing the single Add Component button. I sort of thought clicking on a button was in everyone's knowledge background, but apparently I need to explain how I did it.

  1. I selected the transform,
  2. selected Add Component, and
  3. clicked on it (the Add Component button).

I tried using the Add Component before writing the first time. Nothing happened.

Repeatedly trying grossly wrong procedures is not efficient. What is efficient is to try a process that you think should work once, keeping careful notes.

Before writing the first time I tried clicking the Add Component button (twice), and apparently nothing happened. No centroids appeared (like they would have in M8). No, I was not expecting to have a layer added to the map. I was expecting to have the specific centroids added to the current layer in the map. Nothing appeared to happen because I did forget that the component added was added to the project and not directly to the open drawing as I was expecting. When I click Update Component, the parcels morph to centroids in the current layer of the map. So that's where I expected to see stuff happening. When I click Add Component, they go somewhere else, as TAL so succinctly explained. After I was reminded that the component was added as a table and drawing to the project, that's when I found the new component and copy 2 of the new component. So I had tried the proper/efficient process twice before looking for a grossly inefficient way to make centroids appear in the map. I was not expecting M9 to do the new layer. Perhaps there is room to improve that Add Component button such that it gives a choice to add a new layer to the project and/or add it to the map.

There is just something about poking at individual objects in a sea of 32,000 that smacks of inefficiency.

Walk a mile in my shoes. In M8 with all the labels visible, my county is a fuzzy sea of blackness. I have experimented with making the labels visible at certain altitudes, but that is more work and still leads to a cluttered map as I zoom down. With the labels on I can't see where to zoom in to find an area I might be looking for. With the labels turned off it looks like a map of clean parcels. When someone comes to me and asks who owns the 300-acre property behind theirs, I'm not going to write SQL to find it. I zoom in and click on it to make it active. Once selected the label changes from transparent to black (or white or yellow, depending on the theme). The theme tool is so easy to change I will change it to recolor the text depending on the background I have turned on. When I want to see all the property owned by that person, I will use the search tools for 'owner name > starting with' and type the last name. That's about the extent of the automation I've needed in M8.

As for the limitations of my questions: I ask them in a targeted way hoping the answer will provide that one precious missing step I can use to extrapolate my base of knowledge to make this program work. Notwithstanding the premise for the original post here, I usually don't need to reread stuff I think I already know. The questions I ask are hopefully 5 or 10 steps into a process. Now there are times when I don't know enough about what I'm asking and I will still ask a limited question - I did not think this was one of those times. But with the two Geoms solution, I see that creating pokable labels is way beyond where I can go right now. I sort of feel like a ping pong ball that just got swatted into a basketball court. I am definitely in the wrong place, now.

Dimitri


7,413 post(s)
#20-Jan-18 18:11

But with the two Geoms solution, I see that creating pokable labels is way beyond where I can go right now. I sort of feel like a ping pong ball that just got swatted into a basketball court. I am definitely in the wrong place, now.

I can see that. But that's what happens when you try to use any sophisticated program without first taking the time to learn the basics of the program and how to operate it.

To learn Release 9 you need to invest a solid day or two of fairly intense learning. Skip that and you only spend far more time thrashing around with tons of frustration that people who learned the basics have completely avoided.

Start with the first topics and read them in order - actually read them without skimming - at least through the Getting Started and Basics books. Mix in careful reading of Examples topics and repeat examples yourself.

Beginners at GIS do well with that approach because they don't know any better than to follow instructions, so quite often they actually do read the topics as recommended above. Experienced GIS people sometimes have a far harder time of it because they won't invest into that initial learning phase. Instead, they try to reasonby analogy from what is already familiar to them, augmented by occasionally diving into various topics that seem relevant.

That may not be a bad approach for a system which is very similar to what is already well-known, but it is a terrible approach for something genuinely new. Ifyou fail to learn the basics, yes, for sure you will feel totally lost and confused because not knowing very basic, simple things causes a cascade of frustrations that prevent effective workflow. What is totally easy and fast for somebody who learned the basics first becomes a hall of mirrors and endless frustration for you.

Sure! Why didn't I think of using two Geoms?

That you've missed the opportunity to accomplish sophisticated things in a snap is not something to mock, it is something to correct by investing into your own education. Investing into your own skills is never a waste of time. The most valuable asset you will ever own is your own education and skills.

That you didn't think of using two geoms is not a surprise given you're a beginner, you didn't read the topics that tell you want a geom is, and you haven't read the Example topics where more advanced uses of geoms, such as having more than one geom field, are demonstrated. No problem, as that's the sort of thing you can learn by coming to this forum.

Because I have an extremely limited idea of what a geom is. It seems to be an amorphous container for all the secret graphical intrinsics I can no longer see. There are tons of references to geom in the voluminous manual, but a quick survey of them did not reveal a helpful definition.

Hopping around with a quick survey of what topics you think might be relevant is not as effective for learning as investing the time to learn the basics in the first place. All those tons of references assume you've read the basics. They don't repeat those basics, that defined and described the thing in the beginning.

Geoms are simple and are well defined using clear language in the Tables topic and in the Drawings topic found in the Basics chapter. The Tables topic and the Drawings topic are almost painfully explicit, taking nothing for granted about a reader's understanding or prior experience.

Examples:

The shape and locations of objects in drawingsare stored in geometryfields in a table, using a Manifold data typecalled a geom.

and

Each point, line or area is a recordin a table from which the drawing draws data. Each such record at least has a geometryfield that specifies the location and shape of the point, line or area, and each such record may in addition have other fieldsthat provide additional data for each object. [...] By geometrydata we mean the collection of coordinates that specify the location of a point (a small collection, just one coordinate...), the shape and location of a line, or the shape and location of an area.

If you missed the above,then sure, you might come away with thinking a geom is some amorphous container and encounter totally unnecessary frustration. The missed opportunity to learn what a geom is at the beginning cannot be recovered so easily by browsing all the additional discussion that mentions geoms based on the assumption you know what those are.

I get no credit for the 15 step process I innumerated, but I catch crap for not going into the details of pushing the single Add Component button. I sort of thought clicking on a button was in everyone's knowledge background, but apparently I need to explain how I did it.

  1. I selected the transform,
  2. selected Add Component, and
  3. clicked on it (the Add Component button).

You're not being fair, neither to yourself nor to others.

You did get credit for the 15 step process you enumerated., because your enumeration of the 15 step process made it clear you were taking a wildly wrong approach that indicated you were missing all sorts of basic notions. Providing a simple, five second process is one way to try to show you that learning the basics really does enable you to use simple and fast procedures instead of wildly wrong, time-wasting stuff.

As to clicking buttons, your sarcasm appears to have missed the point that blind clicking of buttons without having first learned what those buttons do and how to use them sensibly is not the way to getting it right the first time, saving time and frustration.

Consider your three step process and the information you did not provide:

1. What transform? What was the window that was open? Was it a drawing, a table, a labels component, an image? A map? What layer? What was the schema for the table?

2. What were the options you used? What did you intend? What did you expect?

3. What did you do with the components that appeared in the Project pane?

If you know what the thing does (well documented) the above are simple questions that for the most part take care of themselves during informed workflow. If all of this is alien for lack of learning, well, sure they may seem confusing.

Walk a mile in my shoes.

Happy to do that to help you find your way, but to walk a mile in your shoes I have to know where you are, in what direction that mile is and what the shoes are you are wearing. Start with an overall description of your task. Then we can help.

It sounds like your task is given a parcel ID to see the parcel. Fine... that's the task to focus on.

I ask them in a targeted way hoping the answer will provide that one precious missing step I can use to extrapolate my base of knowledge to make this program work. Notwithstanding the premise for the original post here, I usually don't need to reread stuff I think I already know. The questions I ask are hopefully 5 or 10 steps into a process.

That's not going to work in a situation where your approach is wildly wrong because you never learned the basics of the thing. Trying to fix step 6 or 11 when the first step took you 180 degrees in the wrong direction cannot give you a solution. It's like getting on a forum asking for an explanation how to shift from P for Park to D for Drive so you can get moving without mentioning you are on a sailboat and for some reason you think there should be a shift lever to operate the sails.

It sounds like your end task is that in a parcels layer you want to see info on a specific parcel. You want to do that with 9 without spending the time to learn the thing from the very beginning.

Your approach, therefore, is to reason by analogy to what you know in 8. So, you're asking questions based on presumption that stuff in 9 works close enough to 8 that you can wing it without really making the effort to first learn 9. To zero in on the parcel you apparently want floating labels to guide you, labels that switch on and off.

But 9 is a very different product, for all the right reasons, than 8, so reasoning by analogy with 8 takes you into wildly wrong workflow. Doing that just causes frustration and takes a hundred times longer to get nowhere.

Learn 9 and use the simple stuff in 9. For example, if you want to find a parcel you don't need SQL. That's what a simple Ctrl-F (for Find, a pretty universal shortcut that is easy to remember) is for. Leave a table open so you can use it when you want. Find the parcel record with a Ctrl F, select it, and you get a bright red dot in the map in what otherwise is a sea of densely packed black. Easy to zoom to that. Don't like the table presentation of attributes? No problem. Alt-click the parcel to see fields in Record - values format.

Now, since you haven't really told us all the ins and outs of what you want to do the above, of course, is not the same comprehensive, everything you might want to do spelled out description. It's just an example that you don't need to use SQL to find a parcel.

But don't mock SQL when for lack of a few hours learning SQL you waste hundreds of hours doing things manually that you could have done in seconds. The ability to use SQL is cool because you can use queries to automate routine tasks. It's a huge time saver.

It's also cool because tiny, really simple snippets of SQL can help you build simple layers and other helpful infrastructure that let you put together a project to save yourself time and hassles. For example, I don't know if having guide layers is a good idea for your task. But if it is, may as well do them right, and SQL can help you create them very quickly and effectively. For example, if parcel numbers have some sort of regional ordering you could create groups of parcels with the same leading edge of most significant numbers, put a dot there and make a label. Then you'd have a great layer for zoomed out guidance.

Like I say, depending on what you want to do almost always there is an easier, better way of doing that in 9 than in 8. But you must commit to learning 9, master the basics, and then use 9 methods to make it easy.

KlausDE

6,410 post(s)
#20-Jan-18 18:48

Given people are as they are we should have two more help topics in Getting Started with links in prominent position in Introduction. These are

  1. What is different for Manifold 8 users
  2. What is different for Arc... users

They should only stress the philosophy of 'everything is a table' and point at basically different concepts.

We must consider that experienced GIS users are bored by the 'Manifold work cycle' in Getting started and tend to close help at this point.

The only reason for these help topics is to capture the impatient and all knowing GIS users and lead them to help topic Basics. This topics should not exceed 40 lines of text. And to take the already knowing serious it could do without images


Do you really want to ruin economy only to save the planet?

tjhb
10,094 post(s)
#21-Jan-18 07:56

I agree in principle with Klaus.

I would also agree that in principle, the current Introduction topic covers most of that ground.

The main problem, in my opinion, is that it does so wearing the wrong hat.

Currently the Introduction topic is wearing a Marketing hat.

I would rather see a Peer-to-Peer hat worn here—which I suspect Klaus is also getting at, in an only slightly different way.

(I tend to think there should be almost no marketing in the manual at all.*)

tjhb
10,094 post(s)
#21-Jan-18 08:19

(*I.e. that users should be given facts here, with few adjectives and adverbs, and left to draw their own conclusions. That still functions well as marketing; I personally think better.)

dchall8
1,008 post(s)
#22-Jan-18 18:05

I used M4-8 since 2004 with little to no change in the user interface. I think I speak for a lot of us in saying that the M9 change in the UI comes as a surprise. I still realize that it is a work in progress, but everything you have said repeats the idea that it will not work like M8. Furthermore knowing SQL is going to be basic to doing tasks which were done with the click of a button or two in M8. No matter how much I want the SQL rqmt to change, it will not. Now you're going to tell me that M9 is not doing the same thing as a few clicks in M8, because I have the ability to do so much more in the rich M9 environment, but you're playing with words. I like having the ability to Ctrl-Alt-click a parcel and immediately move vertices around.

If M9 is to be a GIS for elite SQL users, with a completely rewritten UI, which does new things that the previous version didn't do, then why confuse us by giving it the same name as a much different program? Everyone knows that Notepad is an elementary word processor with a less sophisticated set of features than Microsoft Word. Microsoft could have named both programs Notepad.

It seems like SQL is going to be a prerequisite for M9 simply because the developers have a sophisticated knowledge of SQL. That may not be true, but that is what it appears to be. At this point it seems the only benefit to M9 for me is the launch and save speed. After that, with SQL being a requirement, then everything slows down and the start/save time savings is lost. I could use the ability to import LAS files to visualize clouds in M9, but M9 does not export points to M8, so there ends the utility for me.

There is so much about M9 that I do not like; I need to stop beating my head against it. Then the beatings from this forum will stop, too. Having said that...

Geoms are simple and are well defined using clear language in the Tables topic and in the Drawings topic found in the Basics chapter.

...but if you don't know where to look, and the Contents don't tell you, then you are reduced to using the Index part of the help. There are nearly 100 entries for geom and more than 100 more for geomxxx entries. And if you click, say, entry number 63 and read it, then you come back and have to remember that you clicked 63. I don't remember the number I clicked. I would take notes but I am optimistic I'll find the full answer and not have to return to the entry list. So I start again at entry 1 and continue through until I don't recognize what I read.

At least I know how to create centroids to use to make labels (something M8 did quietly in the background).

adamw


10,447 post(s)
#23-Jan-18 13:30

Furthermore knowing SQL is going to be basic to doing tasks which were done with the click of a button or two in M8.

Maybe I am reading it wrong, but this isn't the idea.

The idea is that you will have easy buttons just like in Manifold 8, sometimes easier. Frequently a bit different, but easy. And that on top of that, if you want to do more, it will be easy to go from buttons to SQL. Because the UI tools write queries for you, because the query engine provides high-level functions and constructs in addition to low-level ones, and because there are no pieces of data internal to the UI that you cannot access via SQL.

No requirement to use SQL. It is just an option.

There is so much about M9 that I do not like; I need to stop beating my head against it.

We would be very grateful for the specifics.

Regarding this thread, suppose Add Component added a new component as a layer (which you suggested in one of the posts and which we were planning to do as well). Should we do anything else?

Dimitri


7,413 post(s)
#23-Jan-18 16:53

everything you have said repeats the idea that it will not work like M8.

Yes. That's been the message since the beginning. It is a very good thing that 9 is a different product than 8. For example, 8 is very slow at big images. We don't want that in 9.

Furthermore knowing SQL is going to be basicto doing tasks which were done with the click of a button or two in M8.

Nonsense. 9 has endless point and click facilities, like hundreds of transform templates. People who have read the documentation and who have learned how to use them tend to get along fine without SQL.

It is certainly true that skilled users will often choose to use SQL instead of point-and-click methods because they can craft their SQL to do in a single line what might take a sequence of point-and-click operations. That's true of SQL in 8, by the way, not just of SQL in 9.

SQL in 9 is even stronger and better than in 8, so it is no surprise that skilled users in 9 will more frequently choose to use SQL. That skilled users in 9 enjoy the convenience and power of SQL doesn't mean they don't know how to point and click, it just means they like the convenience of SQL. It is good that people often have the option to either point and click or to use SQL. That is true of 8 and it is true of 9.

I like having the ability to Ctrl-Alt-click a parcel and immediately move vertices around.

Same in 9. See the Editing Drawingsin the Basics chapter. Atl-click a parcel and immediately move vertices around to your heart's content. No SQL required. Just drag with the mouse and, like magic, the vertices move to where you want them to be. They even respond to snaps.

why confuse us by giving it the same name as a much different program?

Anybody who thinks 9 is the same as 8 hasn't been paying attention. It is not a name issue. Release 9 is a famously different name than Release 8.

From the very first mentions of 9, absolutely every source on 9 has emphasized it is a totally new thing that is very different from 8, for all the right reasons. In many thousands of posts and discussions as 9 has emerged from Radian Studio evolving into Manifold Future into the 9 pre-release, the loud and clear message has always been that 9 is totally new, it is very different from 8 and that you cannot learn it by trying to reason by analogy with 8.

It seems like SQL is going to be a prerequisite [...] with SQL being a requirement,

Nonsense. Most people never touch SQL with 9. It's a very useful option, not a requirement for most people. They just point and click away using hundreds of, well, point and click facilities, like the Transform panel, the Select panel and others.

but if you don't know where to look, and the Contents don't tell you, then you are reduced to using the Index part of the help.

The above is the complaint of somebody who has refused to follow the instructions on how to learn 9. Yes, if you refuse to follow the instructions and instead try a Rube Goldberg, bass-ackwards way of trying to learn the thing, in truth the process will be much more difficult and frustrating. It's crazy to try to learn 9 by hunting and pecking using the Index. Instead, learn 9 by following the instructions.

Let me try an analogy. If you are going to France for a few months to live in a country town where nobody speaks English, and you don't speak any French, well, it would be wise of you to learn a bit of French language.

A good introductory French text will set you up in a series of introductory chapters to get the basic ideas under your belt with a beginning vocabulary. That's a good way to learn. Compare that to two really awful and inefficient ways to learn:

1. "French should be like English, or German, so I'll just try winging it on what I know of English or German" - That's not going to work. If you are in the hardware store and you need to buy a lightbulb, asking the French clerk for a lightbulb in English or German isn't going to work. It doesn't work any better if you try saying the same English or German words more slowly, or louder. It's still not French.

2. Deciding you don't have time to read the text, but that once you are in France you can look up words you see on French signs in the Index and learn from that. That's a better approach than beating up a French clerk with German words, but it will take you about a hundred times longer to learn French by hunting and pecking looking up words in the Index. You should have just read the introductory chapters in order, done the exercises, and learned the vocabulary so you can build bigger steps upon a good foundation.

To learn 9, follow the instructions for learning 9. This begins with the Read Me First topic.

The reason the Read Me First topic is called "Read Me First" is because that is where you should start.

At the end of the Read Me First topic it says...

Let's go! Start with the Introductiontopic.

It doesn't say "stop reading now, and then thrash around trying to find what you didn't learn by using the Index." It says:

Let's go! Start with the Introductiontopic.

So that's what you should read next. The Introduction topic ends with...

Read these Topics Next

Getting Started

User Interface Basics

You can see where I'm going with this. At no point do the introductory topics tell you to blow off reading the user manual. Each of the initial topics points the way forward. The next step is a big one, since the Getting Started topic tells you what you must do by way of further learning:

Learn

  • Read the topics in the Basics chapter.

  • Master the main dialogs:

All that is rock solid advice that will save tons of time and help avoid unnecessary frustration. For example:

  • Read the topics in the Basics chapter.

The above are big steps, but totally useful steps. Reading the Basics topics saves a lot of frustration and helps you get going quickly with much less effort. No need to pour through the Index.

The right way to learn 9 is simple: start with the Read Me First topic and then steadily, thoroughly, attentively read each and every topic right up through the end of the Basics chapter. Drill down into the topics and the examples those topics in turn recommend. That will clear up a lot.

That's a lot of reading and if you want the benefit of the technology in the pre-release you necessarily will encounter some version skew working with Release 9 Edge, which you should be using. The documentation cannot keep up with how fast the thing is evolving.

But the basics have caught up with the builds, with the exception of images, where the new conversion to a Style panel from the old Edit - Style dialog is introducing numerous new, ahem, point and click controls. But it sounds like you have a vector focus so you shouldn't have any issues with that.

Don't ever cut corners by skimming the documentation thinking you can work 9 as if it were 8. It's not, and it's not 8 for all the right reasons. If you don't feel in your bones why those reasons are the right reasons, you are missing very important understanding of both 8 and 9.

I am proud to say I played a role in the design of 8. Like all Manifold stuff 8 was a team effort so I cannot say I did more than make a contribution, but that's enough for me to be able to honestly say i can take pride in 8. 8 is one heck of a great product.

9 is profoundly better. I say that even knowing 9 as it is today is not remotely close to final form. If you like 8, you're cheating nobody but yourself by not giving yourself the chance to learn 9 as 9, not through some reflection of what you know and expect from 8.

Trust me on that as one of the guys who made 8 and who now has helped make 9. 9 is worth the effort to learn right. If you think all 9 has to offer is quick startup and similar speed, once you get your head around it you'll be pleased to see that the way 9 really saves you time is by making more efficient and way more maintainable workflows possible than in 8, along with a zillion other advances.

Graeme

990 post(s)
#24-Jan-18 03:33

Same in 9. See the Editing Drawingsin the Basics chapter. Atl-click a parcel and immediately move vertices around to your heart's content. No SQL required. Just drag with the mouse and, like magic, the vertices move to where you want them to be. They even respond to snaps.

There doesn't appear to be the equivalent of M8 "shared edit", where coincident coordinates in touching objects will all move when feature is turned on. We find that functionality essential for manually editing area objects. Any indication of (hopefully) when for 9?

I can't find reference to "snap" in the edit drawing object sense in the documentation - that is snapping to the same location as another coordinate in the same or other drawing object. Could you please indulge me with a link to the relevant help topic?

Dimitri


7,413 post(s)
#24-Jan-18 11:29

There doesn't appear to be the equivalent of M8 "shared edit", where coincident coordinates in touching objects will all move when feature is turned on

Correct. I note that with snapping it is easy to move coincident vertices in, say, adjacent objects by alt-clicking the adjacent object and moving the other vertex as you moved the original one.

Could you please indulge me with a link to the relevant help topic?

Sure, that's covered in the recommended video, in the Videos section of the Editing Drawings topic. See the Future Tour Part 2 Editing video.

A note...

Any indication of (hopefully) when for 9?

... when and if people prioritize it. Then it will appear. If nobody gives a hoot about it enough to have sent in suggestions, well, then other things which people care about more will get done first. See also the FAQ for an overall guide to what is planned for 9.

Keep in mind it takes time for an engineer to look into a highly complex process involving many thousands of tasks to issue a prediction when something may or may not appear, so let's reserve that for maximum priorities. If something is important to you, do not hesitate to send it in as a suggestion. If something was not sufficiently important for that, don't ask an engineer to burn time on getting you a prediction.

That folks like Adam spend hours on this forum in addition to their day jobs is very cool, but every moment they spend looking up predictions is a moment they are not spending implementing one of those features we all want. I, for one, would like to allow them some time for sleep, nutrition, a bit of exercise and all those other annoying activities which interfere with life (that is, doing Manifold...). :-)

As always, to learn how to best ensure that your tastes, preferences and priorities have maximum impact on the development of the product, read and apply the advice in the Suggestions page.

Dimitri


7,413 post(s)
#24-Jan-18 19:12

I note that with snapping it is easy to move coincident vertices in, say, adjacent objects by alt-clicking the adjacent object and moving the other vertex as you moved the original one.

Correction: The above is a mistake on my part, as I don't see that snaps are turned on in the published build when editing an alt-clicked object. They should be. I'll look into this.

adamw


10,447 post(s)
#23-Jan-18 13:22

Just in case:

Perhaps there is room to improve that Add Component button such that it gives a choice to add a new layer to the project and/or add it to the map.

We have been considering doing this, perhaps automatically. We didn't do it yet, because there are multiple ways to go when the map is read-only and we have to choose the best one.

tjhb
10,094 post(s)
#23-Jan-18 13:31

It would be intuitive (and not annoying). I.e. if a map is open and active when a transform is run, for the new component to be added to the map automatically. Perhaps immediately above the active layer? The new layer not automatically activated.

(If the map is read-only, but the project is not, just silently do what is possible? Add new component to project root, leave map alone.)

Dimitri


7,413 post(s)
#23-Jan-18 17:20

The action button settings are a fork in the road for what you want the thing to do: change the data in place or put the changes into a new component. That's a simple, clear, binary choice that should not be complicated by trying to combine different functions into the control.

Likewise, deciding what you want in a window is a simple thing: Want a layer? Drag and drop it into the window. Done.

That simple act automatically deals with issues like not being able to drop into the window because it is read-only, and it deals with issues like maybe the window not being a map but a drawing and so on. The less that stuff happens with the system trying to guess its way through to what you really mean, sometimes adding, sometimes not, the better.

If somebody doesn't know how to use a control for lack of reading the manual, adding more complexity isn't going to solve that issue. It will be just more confusing because sometimes the control adds stuff and sometimes it doesn't. Such inconsistency is not good even for people who do read the doc.

Maybe a different way: Yes, RTFM is good but anything that can be done for captions to be self-documenting is also good. So maybe the caption "Add Component" should be changed to "New Component" or "Create Component". That could be a stronger tip something new will be created in the Project pane.

tjhb
10,094 post(s)
#23-Jan-18 19:27

That is all true. And the user interface for 9 is intentionally very pure.

But design purity is only as good as its ability to allow occasional exceptions.

Maybe this is not one, but it would be intuitive and do no harm.

It would be most useful when only the Contents pane is visible, the Project pane inactive, which is (rightly) the default configuration.

In that case pressing the "Add component" button gives no visual feedback at all for a moderately-sized task. To put it another way, 9 is way too fast. Some (any) visual feedback is all that's wanted I think.

Yes, a bigger screen, or multiple screens, is possibly a better answer--but the UI can't target the perfect case.

I think "Create component" would be a helpful improvement over "Add component". (The user is not encouraged in thinking "Add component to the map", if this will not happen.) Amazing how subtle these things can be! To me that seems a good call, it might be the only tweak needed, though I ensuring that there is always some visual feedback might be even better.

Mike Pelletier

2,122 post(s)
#23-Jan-18 23:22

Nice thoughts Tim. For the visual feedback, how about a brief note appear in the bottom bar.

Manifold User Community Use Agreement Copyright (C) 2007-2021 Manifold Software Limited. All rights reserved.