TheSwamp
CAD Forums => Vertically Challenged => Land Lubber / Geographically Positioned => Topic started by: Dinosaur on November 30, 2006, 06:55:07 PM
-
The time has now come to give Civil 3D a real test. This is our newest project . . . 163 acres with a mix of townhomes and traditional single family attached development over at least 5 phases. Elevations throughout the site range from 920 to 995 with 3 natural drainage basins. The current plan is for ALL work to be done using Civil 3D 2007 or its successors along with Hydraflow for the storm water design calculations.
-
Good luck and have fun! I wish I could land a project like that!
Keep us posted with your progress.
-
That should be right in the sweet spot of the software.
-
That should be right in the sweet spot of the software.
I hope so, because this is exactly the type of project we need it to do. I have not heard of any other projects of this scope being tried with Civil 3D yet, so any war stories or advice will be most welcome. I am more worried about the phasing than anything else because it is likely by the third phase we could have as-built, staking and design work active at the same time.
-
I'm curious as to how easy it is to do the phasing.
Land Desktop wasn't really designed for phasing. We managed to hack our way through using the PROJECTNAME feature, and that's how we've been dealing with long-term, phased projects in Land Desktop. It doesn't work very well, but it's been better than nothing.
Does C3D have any special features for handling long-term phased projects? I haven't seen any yet, but I haven't really looked. Do Data Shortcuts honor the PROJECTNAME settings?
-
Me thinks that stacking the phases as sites would be the start point.
If the 'as-built' data is used to update the live model, then it would
be possible to overlay this information on top of the archived submitted
files to produce the 'as-built' documentation set. All the while keeping the
design model as correct as possible.
-
We set up our LDT multi-phase projects by just creating a new project folder with copies of everything from the previous phase. There was usually enough time between the latest approval and startup for the next phase we only had to adjust to the asbuilts at some point during the design phase. I am not sure how well this procedure will stand with Civil 3D but that will be our initial attempt. There may be some things that can be handled with sites but the points and surfaces are not site specific. Perhaps point groups and grading groups will also be helpful but this is all uncharted territory for us.
-
Use Vault and you can push and pull things between the phases by having them as different projects and simply referencing across Project frames. You can create data references across projects, making phasing considerably simpler than it was in LDT or C3D 2006.
Vault IS the way you want to go. Go to AU Online if you're on subscription. Get my lab, Dan Philbrick's Project Management class, and Anthony Governanti's Keep the Data and Get that Sheet out of Here class (Sheet creation basically.)
If you're taking a plunge, take a plunge and do it right.
-
Vault IS the way you want to go. Go to AU Online if you're on subscription. Get my lab, Dan Philbrick's Project Management class, and Anthony Governanti's Keep the Data and Get that Sheet out of Here class (Sheet creation basically.)
If you're taking a plunge, take a plunge and do it right.
The problem with Vault is that it involves setting up a server, getting SQL server running (the OOTB Vault is worthless), and all that mess. Quite a lot to handle when everyone's simply trying to use Civil-3D.
Civil-3D SERIOUSLY needs ways to do this stuff that don't involve Vault. Then, once people have an idea how to use Civil-3D, they can worry about adding Vault in for Document Management. This idea of trying to force Vault (a version control utility) to handel C3D's Project Management features is a terrible idea on many levels.
-
Vault IS the way you want to go. Go to AU Online if you're on subscription. Get my lab, Dan Philbrick's Project Management class, and Anthony Governanti's Keep the Data and Get that Sheet out of Here class (Sheet creation basically.)
If you're taking a plunge, take a plunge and do it right.
The problem with Vault is that it involves setting up a server, getting SQL server running (the OOTB Vault is worthless), and all that mess. Quite a lot to handle when everyone's simply trying to use Civil-3D.
This simply isn't true Sinc. I run Vault on a VMWare XP box inside my laptop WHILE I'm running C3D for my daily work. It's not ideal, but it works. Most clients can use a workstation they have lying about to handle the Vault server in terms of hardware and OS requirements, and the SQL Express 2005 upgrade is free to take you to from the MSDE version that ships.
Beyond installation of my original Vault, plus the SP that was released last week, I have NEVER touched my SQL server. JP doesn't like Vault bexause of the hardwrae, but ask him how much maintenance he's actually had to do at the SQL level. I'm pretty damned sure it's none. I have another client that is using his old LAPTOP with an attached USB2 drive for his server. The requirements simply aren't that arduous, especially at a pilot project level.
You wouldn't ask someone to consider how to do site layout without a basic understanding of site topography, why would you ask them to design and create a dynamic 3D model without understanding at some basic level how it all goes together.
I would seriously be happy to work with or talk on the phone with anyone who wants to go into this. Once you see how Vault can be levereaged and the actual disadvantages of Shortcuts, I think you'll be convinced to give Vault a very serious runout.
-
I would (and have) seriously given thought to the idea of using Vault. My problem is that We are still literally teetering on the edge with Civil 3D. If I have this project in a Vault and ANYTHING goes wrong such as an accidental overwrite of data when there are multiple users, the entire idea of using Civil 3D may be tossed. At minimum, John will easily hear the resulting uproar from his far corner of the city. My engineer is quite undisciplined in drafting mode at best but will be actively involved within the model most of the time and my supporting staff is as raw as you can get. We at least have a very good backup procedure that can restore any daily state since it was implemented, but a day is too much to loose. I know solid training is the best insurance, but that route has been blocked for 18 months with no sign that it will change. A disaster MAY change that, but would more likely mark the end of the Civil 3D era here.
-
>> I have NEVER touched my SQL server. JP doesn't like Vault because of the hardware, but ask him how much maintenance he's actually had to do at the SQL level. I'm pretty damned sure it's none.<<
True.
I don't like it based on drive space vs value delivered. Among other things.
But the hardware requirements are NOT robust in a smaller environment.
And James is correct that Jr. SQL will preform quite well in the 20 to 30 user range.
-
Is there any feasible way to run Vault, share the data with up to 5 users, merge the vault into our existing backup routine
AND
do all of this WITHOUT a Windows server?
We put a $5000 Linux server on line 2 months ago to replace our old Windows server than was constant problems and IT wants no part of having any kind of windows server interacting with it. Our existing workstations are AMD 3500 or above cpu's on XP PRO with either 72mb Raptors or the Seagate parallel write drives. I have to agree with him that network stability has been significantly improved with this change. He is also claiming that if a windows server for Vault was introduced into our network, the necessary changes would cause noticeable decrease in performance. I don't know about this, but that is the line in the sand . . . without meeting the above criteria there will be no Vault.
-
You wouldn't ask someone to consider how to do site layout without a basic understanding of site topography, why would you ask them to design and create a dynamic 3D model without understanding at some basic level how it all goes together.
Exactly. That's why Autodesk should fix the problem correctly, rather than trying to band-aid together a solution.
The main problem is that they are confusing three entirely seperate things:
- Project management. By Project Management, I mean the grouping of related drawings into a coherent whole. The Sheet Set Manager is part of this, but not all of it - for example, it should be possible to have many different Sheet Sets inside the same project. Fields should be tied to the project, as well, so that project data like Project Name and so forth could be set once in the project, and then used in any drawing after that without having to set it anywhere else. In other words, if the Project Name should change, the user should only have to change it in one place, and that change should immediately be reflected in every DWG in every Sheet Set in the project. It should be possible to configure "jumping points" based on the Project home directory. For example, when XREF'ing a drawing, it should be possible to jump right to the Project's drawing directory, rather than always needing to browse from the location of the last XREF. (There's a TON of other ways that good Project Management would eliminate most of the arduous browsing generally required by Autocad today.) It should be possible to link multiple Projects, and easily maintain the links when projects get moved or archived. But it should go even deeper. For example, it should be possible to specify which set of Standards are to be used for all drawings in this particular project. Then, by default, all drawings are created with the correct template, in the correct units, and the right set of Tool Palettes will be visible (i.e., if you're working in an Imperial drawing, you only see Imperial Tool Palettes, and not a bunch of Metric ones), the drawing references the right ACAD.LIN file, etc. (This is particularly important for people who have to do work for multiple government agencies, each of which uses a different "Standard" that they require all their DWG files to follow.) There's a lot more, but this is the basic idea.
- Concurrent, shared user access to Model Data (aka Model Data Management). This seems most people actually mean when they say "Project Management" (probably because Softdesk and Land Desktop call their model data "projects"). It is controlling the creation, storage, and access of the logical information of the model, which can be shown in various ways inside of drawings. When an element of the model is created or modified in any drawing in the project, that data should be pushed to the model. When changes get pushed to the model, a notification should be sent to other users who are currently accessing the data. Locking of model data should occur only as-needed - data should not be locked simply because I open a drawing that contains a reference, or because I am using the reference in some read-only way. This locking should occur automatically, by the software, with no need of concious thought by the user.
- Version Control and Document Management. Software Developers have been using this for years. All data actually resides inside a "vault". In order to change any data, a user must explicitly "check out" the data. Then, once changes are done, the user must "check in" the data. At any time, it is possible to extract the data as it existed at any previous moment in time. There is also a transaction log maintained, containing a log of what files were changed, by whom, and any note that person might choose to leave.
Now the first item is something that would be incredibly useful. But it is something that should be included in VANILLA AUTOCAD, not bandaged onto the product through the Vault. And Autodesk does not seem to feel that it is useful, anyway. Either that, or they think it would be too useful, and too powerful to put in Vanilla Autocad. I admit that Autodesk's motives are often obscure, but it's a complete mystery as to why they've gone so many years without putting PM into Vanilla, where it belongs. They added lukewarm PM to Land Desktop, and they added relatively decent PM to ADT, but for C3D we get the Vault? Huh? Why not solve the problem in Vanilla Autocad, so that it is solved in ALL THE VERTICALS, too?
The second is basically what Autodesk is trying to use the Vault to accomplish. But the Vault is not really designed for this task - it's the sort of task that is rightfully done by a transactional database that contains the model information. Instead of doing that, Autodesk is placing the model information in drawings, and then using the Vault to control access to the drawings. That's not a correct solution. Their method forces the user to think a lot about the layout of the project, and do an awful lot of work manually. The user has to decide with drawings to put model elements in. The user has to check out drawings and check in drawings in order to do anything. Nothing happens automatically. It works, but it is not a user-friendly experience, and doesn't work anywhere as well as the so-called "optimistic locking" strategy that a true multi-user, concurrent-access database could manage.
The third is actually what the Vault was designed for. But in Civil work, it really isn't necessary. Most companies know how to make backups of projects at key points, and they can always pull out the backup if they need to. They don't need Vault for this. And the accountability trail isn't really a reason to use Vault, either. If it is important to track who did what and when, then a simple log can take care of that. The Vault is overkill. I don't know about other companies, but in four years in business, we've never wanted or needed the accountability feature of the Vault, and we've never wanted or needed to go back to an earlier version of the project. That type of functionality is superfluous for us.
So basically, Autodesk is failing to address one part, addressing one part incorrectly, and fully-addressing the part that no one cares about. And they wonder why everyone is complaining? :|
I didn't realize there was a SQL Express. That might solve my main issue. The included MSDE version may suffice for you, being by yourself like that, but in a full office environment, it doesn't cut it. I actually have some desire to use Vault. I think it might be a real solution to our two-office conundrum (we have to share projects between offices, which has been a management headache in Land Desktop with no Vault.) But I would much prefer to add the Vault later, after we're comfortable with C3D.
Let's see... All we have to do is set up a SQL Server and the Vault, and figure out how to manage and maintain it, and train all our users in how to use the Vault when learning to use C3D is taxing them to their limits, all without losing anything or causing significantly more down-time than the transition to C3D is already causing... Yuk! That's not the sort of task that anyone should be forced to do, if the idea is to make a smooth transition... Good thing I have a software engineering background - the guys at this office are great at what they do, but they are definitely not IT or computer professionals. They would never manage this sort of transition without the sort of expertise I bring - if it weren't for me, they'd probably be on Land Desktop for at least a couple more years. Autodesk really thinks this is a reasonable thing to expect from ALL their clients? :ugly:
-
Sinc I'll take a shot at this when I have time to think it over.
Lot of points made, some that I agree with and some I don't.
A little time to digest and I will get back.
Dino, as far as putting your eggs in a Linux server, sorry to say it BUT.
>> He is also claiming that if a windows server for Vault was introduced into our network, the necessary changes would cause noticeable decrease in performance.<<<<
I'm calling B*llsh*t on this one.
-
. . . I'm calling B*llsh*t on this one.
I am of the same basic opinion, but his additional post as eldest son of boss and heir apparent trumps you, me, sinc, James Wedding, Michael Farrell and the collective wisdom from Manchester.
-
Sinc I'll take a shot at this when I have time to think it over.
Lot of points made, some that I agree with and some I don't.
A little time to digest and I will get back.
Well, you have a point, too. Just because Autodesk is horribly mangling the PM and model management doesn't mean we shouldn't use the tools they aremaking available. But it is still a lot easier to add components one at a time. There is no good reason for using the Vault to control access to model data, except probably that someone at Autodesk thought the accountability of the check-in/check-out procedure would provide some value. It doesn't, really. It just adds hassle. (They probably also thought, "Hey, we already have this Vault, you think we could use it as an application server?", not knowing that that was a terrible idea...)
At some point in the process, I'll get the Vault setup on our server, and we'll try a project on it. But I plan on waiting a bit. We have Civil-3D working without the Vault, although this means we get basically no Project support at all from the program - merely the ability to create data shortcuts. However, until we get over all the bugs and gotchas involved here, I don't want to throw Vault into the mix. I have enough to worry about, trying to get an entire office moved to Civil-3D while still learning the program myself, and still maintaining productivity. We can't fail to meet our client's expectations just because we're switching to new software, and there's only so many hours in the day...