Tag Archives: AutoCAD 2018

Fix released for Autodesk subscription licensing error (yes, another one)

Providing yet another entry in the it’ll never happen again file, Autodesk has issued a fix for an AutoCAD 2018 subscription licensing crash.

Thankfully, it’s not compulsory to use the execrable Autodesk desktop app or malware-like Akamai download manager to obtain this fix. There are direct links for the 32-bit and 64-bit fixes.

Lessons we learned from this:

  • AutoCAD doesn’t just phone home once every 30 days to maintain even a long-term a subscription license, it phones home 24 hours later too. And probably every 24 hours after that.
  • Desktop software that relies on the internet to continue working is a terrible idea.
  • Yes, it will happen again.

AutoCAD 2018 for Mac – welcome to twenty years ago

In the past, I’ve described how AutoCAD for Mac was released half-baked (as I predicted) and has remained half-baked ever since.

But wait! Autodesk has proudly announced AutoCAD 2018 for Mac. Skimming through that blog post, I must admit my jaw dropped when I saw some of the new features. This one, for example:

This “new feature” was first provided to AutoCAD users in the 20th century. It was an Express Tool in AutoCAD 2000 (released 1999) and was absorbed into mainstream AutoCAD a few years later. The alias editor goes back even further, to the Release 14 Bonus Tools (1997). That one was absorbed into AutoCAD in 1999. Some of the other new features are also old. Migrating your settings was new back in the century that started without powered flight. Now, not so much.

These features are new to AutoCAD for Mac, of course, and that’s kind of the point. Autodesk is advertising, as new, features that were born before some of the adults who are now using their products.

There are other very important features (e.g. DCL support, essential for LISP compatibility) that date back even longer (Release 12, 1992) and which are still missing from AutoCAD for Mac. That’s right, in some areas AutoCAD for Mac is a quarter of a century behind. And counting.

On the bright side, you do now get access to the pointlessly-changed 2018 DWG format. A couple of features are reasonably new additions, but they represent a small subset of the small number of minor improvements in AutoCAD 2017 and 2018 for Windows. If anything, the rate of improvement of AutoCAD for Mac is lagging behind even the glacial progress of AutoCAD for Windows, despite starting from a much lower base point.

I note with interest that Autodesk’s comparison page is now hiding the detail of the differences between the full product and AutoCAD for Mac. I guess if you have two identically-priced products and one’s missing a bunch of stuff, you might be tempted to hide the fact from your potential customers. This post of mine from last year will give you some idea of what Autodesk’s not telling you about what’s missing from AutoCAD for Mac. Clue: it’s a lot.

Mac users pay full price for their product and deserve much better than this. If you want information on a full-featured “AutoCAD for Mac”, don’t bother looking for it from Autodesk. Try Bricsys instead.

AutoCAD 2018.0.2 arrives

AutoCAD 2018.0.1 is dead, long live 2018.0.2!

Here’s the readme.
Here’s the 64-bit direct link.
Here’s the 32-bit direct link.

This supposedly fixes stuff that 2018.0.1 broke, such as the signed VLX thing. Will this one break other stuff? I guess we’ll find out.

Bloatware – a tale of two installations

In a previous post, I showed that AutoCAD is bloatware by comparing the size of its downloads to that of BricsCAD. Obviously, an application that’s ten times the size it should be is going to cost you a lot of unnecessary bandwidth, download time and drive space. But maybe you don’t care about that. What practical difference does it make?

Well, for one thing, the blimping-out of Autodesk’s former flagship product has a big effect on installation time. Vast and ever-increasing amounts of time are wasted by users of Autodesk products, just waiting for the things to finish installing. But isn’t this just the inevitable price to pay for the functionality provided?

No. Again, BricsCAD proves it.

The installation comparison is shown below. These installations were performed on a mid-range Windows 10 i7 PC with 8 GB RAM. The downloaded files were executed from a local hard drive and the applications were installed to a local SSD. If I needed to enter information manually, I stopped the clock while that was going on. Times are the total elapsed time from commencement in minutes and seconds. More user input was required for the AutoCAD install, but that has not been counted in this comparison. That is, by eliminating the human input stages I’m being kind to Autodesk.

I performed a complete default installation of BricsCAD. In the case of AutoCAD, I turned off the installation of Recap and A360 Desktop to make for a fair comparison, as equivalents are not part of the BricsCAD install and those components are not required by the average CAD user. Everything else was as per default settings.

BricsCAD V17.2 64-bit Windows Installation
Installation Operation
Timestamp
Execute BricsCAD-V17.2.03-1-en_US(x64).msi 0:00
Prompt for questions 0:03
Click Yes for UAC Allow 0:14
Installation complete, start application 0:35
First run startup 0:40
Total time from install to ready to draw
0:40 (100%)

That’s astonishingly fast. Remember, this is an application that is more capable than AutoCAD overall. How does installing AutoCAD itself compare?

Equivalent AutoCAD 2018 64-bit Windows Installation
Installation Operation
Timestamp
Execute AutoCAD_2018_English_Win_64bit_dlm_001_002.sfx.exe 0:00
Self-extractor finishes initializing 2:46
Self-extractor finishes extracting, click Yes for UAC Allow 4:56
Install screen appears, answer questions, start install proper 5:04
Desktop icon appears 8:10
Install complete, restart required 10:21
Restart complete, start application 11:41
Activation begins 12:05
Activation complete 12:21
Close AutoCAD, execute AutoCAD_2018_Product_Help_English_Win_32_64bit_dlm.sfx.exe 12:25
Self-extractor finishes initializing 14:44
Self-extractor finishes extracting, click Yes for UAC Allow 14:54
Install screen appears, answer questions, start install proper 14:57
Offline Help installation complete, execute AutoCAD_2018.0.1_64bit_r2.exe 15:44
2018.0.1 install complete, start AutoCAD 16:50
Second startup complete 17:18
Total time from install to ready to draw
17:18 (2595%)

Installed sizes are roughly 0.5 GB for BricsCAD and 2.4 GB for AutoCAD. It’s hard to be exact because Autodesk likes to perform multiple installs when one is requested and tends to squirrel away various components in a variety of places. Here are the ten(!) new entries in Add or Remove Programs after just the first stage of the AutoCAD install:

OK, so maybe AutoCAD takes 26 times as long as BricsCAD to install. But the AutoCAD installation images are so much prettier than the plain old BricsCAD dialogs! Shall we call it a draw?

No.

Autodesk, you took a real pounding here. Bricsys chewed you up, spat you out, ground the chewings into the dust, set fire to the remains and then put out the fire with bodily fluids. Sorry, but you deserve it. Your installations have been ridiculously slow for years and are getting worse. Installing a vertical product or suite is beyond a joke; it makes even the AutoCAD install look speedy. It’s not good enough.

AutoCAD 2018.0.1 mystery deepens with silent withdrawal

As I mentioned earlier, the release of AutoCAD 2018 was followed almost instantaneously by the first update, 2018.0.1. At the time of writing, there was no official information about this update. Some information was later made available, but questions remained.

Now the update has been silently withdrawn. Go to Autodesk Account > Management > AutoCAD > Downloads > Updates & Add-ons and you will no longer see this:

The infamous Autodesk desktop app also shows no sign of this update. So why has it been withdrawn? Autodesk isn’t saying, but thanks to Jimmy Bergmark, we know that installing the 2018.0.1 update re-introduces a bug from AutoCAD 2016 (pre SP1) where signed VLX files don’t load. This means various 3rd party applications won’t load if the developers have done the Autodesk-recommended right thing by digitally signing their code.

If you’re a developer and want to test your code under the different versions, these direct links still work at the time of writing:

If you’re not sure whether or not you have 2018.0.1 installed, the About command will show you.

You can also check for this under program control by inspecting the system variable _VERNUM. In AutoCAD, it’s “O.49.0.0” before the patch and “O.61.0.0” after. I don’t know about LT, and I don’t know about the situation with verticals. Do they incorporate the 2018.0.1 fixes? How about the VLX bug? Should users who have applied this update uninstall it? Is this going to be done automatically or by Autodesk desktop app? How should users manually revert to the pre-2018.0.1 state if they need to load applications that use signed VLX files?

I think it’s fair to say that Autodesk’s management of this update has been a disaster. This is just one in a long line of AutoCAD update screw-ups going back decades. It proves comprehensively that continuous updates from Autodesk are a non-starter.

Autodesk can’t be trusted avoid breaking things with its updates. It can’t be trusted to effectively communicate about the updates. It can’t be trusted to provide fixes for its broken fixes. It can’t be trusted to provide an automated update mechanism that doesn’t hog your resources or one that works properly.

The AutoCAD 2018 install inflicts the execrable Autodesk desktop app on your systems without asking, which in itself is a betrayal of trust. I recommend you uninstall it immediately after all Autodesk installs. You will need to right-click the app tray icon and use the Exit option before you can uninstall it using Add or Remove Programs.

Autodesk needs your trust to make its continuous update idea work. Autodesk doesn’t have that trust. Autodesk doesn’t come close to deserving it.

Bloatware – a tale of two CAD applications

You may have seen me mention in passing that AutoCAD is bloatware. That’s not just the general grumpy-old-user moan you see from long-term users like me, who can remember when AutoCAD used to fit on one floppy disk.

Yes, programs get bigger over time as new functionality is added and old functionality needs to be retained. Hardware gets bigger, better, faster over time to compensate for that. I get that. Understood.

The AutoCAD bloatware problem is much more than that. AutoCAD is literally ten times the size it needs to be, to provide the functionality it does.

How do I know? BricsCAD proves it. Here’s what I mean.

BricsCAD V17.2 64-bit Windows Download
Downloaded File Size (KB)
BricsCAD-V17.2.03-1-en_US(x64).msi 248,812
Total (1 file) 248,812 (100%)
Equivalent AutoCAD 2018 Downloads
Downloaded File Size (KB)
AutoCAD_2018_English_Win_64bit_dlm_001_002.sfx.exe 2,065,829
AutoCAD_2018_English_Win_64bit_dlm_002_002.sfx.exe 328,277
AutoCAD_2018.0.1_64bit_r2.exe 120,663
AutoCAD_2018_Product_Help_English_Win_32_64bit_dlm.sfx.exe 180,013
Total (4 files) 2,694,782 (1083%)

Which dog is which? They’re both cute, but which would you put your money on in a race?

(Original image: Przykuta)

(Original image: Lisa Cyr)

I’m actually being very generous to Autodesk in this comparison. The two primary AutoCAD download executables alone expand from 2.4 GB to 5.2 GB before install, requiring a total at least 7.6 GB of disk space before we even get to the same ready-to-install point as BricsCAD’s 0.24 GB MSI file.

No, it’s not because BricsCAD is a cut-down application compared with AutoCAD. The opposite is true. Overall, BricsCAD is significantly more feature-rich than AutoCAD. It near-exactly duplicates over 95% of AutoCAD’s functionality and then adds a big slab of its own on top of that. Some of it is in paid-for optional extras, but the code that provides that functionality is still included in the same small download.

This issue isn’t unique to AutoCAD. Super-morbid obesity seems to be standard among Autodesk products. The AutoCAD-based verticals that add a comparable level of functionality to that the BricsCAD download includes are much bigger again!

Anybody care to explain what’s going on here?

AutoCAD 2018.0.1 mystery partially resolved but questions remain

As I mentioned earlier, the release of AutoCAD 2018 was followed almost instantaneously by the first update, 2018.0.1. At the time of writing, there was no official information about this update. Some information is now available, but more questions have arisen.

If, like me, you don’t/won’t/can’t have Autodesk desktop app running on your systems, the only current official way to get at the download is using Autodesk Account (but read the whole of this post before you go there). That’s also how you get at information about the update. Go to Management > AutoCAD > Downloads > Updates & Add-ons. From there, it’s not obvious how to get the information, but it’s under More options.

From there, pick View Details. This will show you the following information (after you pick More):

As you can see, the severity is considered high. If you pick View release notes, you can see the readme, or you can go straight to it if you have this direct link. Here are the fixes described in the readme:

  • Occasional crashes when ending an AutoCAD session using specific API code no longer occur.
  • Publishing annotative multiline attributes no longer results in incorrect annotative scaling.
  • PFB fonts can now be compiled successfully as SHX files.
  • The border of a mask is no longer plotted in PDFs when “Lines Merge” is turned on.

Although it would be easy to have a go at Autodesk for shipping a product that needs fixing within hours of release, that wouldn’t be entirely fair. No software is flawless. Stuff happens, and the sooner fixes are provided to resolve that stuff, the better. So I commend Autodesk for getting this fix out quickly.

That doesn’t mean Autodesk is blameless, though. Read on.

First, the way the information about this update was (or wasn’t) disseminated was sub-optimal. It has required too much prodding and guesswork to get to the point we are now, and we’re still not where we should be.

Next, there’s scant information in the readme. I don’t see any documented way of including this fix in a deployment, for example. That means it’s not possible to create a one-step automated install without resorting to trickery.

Further, this update isn’t available on the main Autodesk site. It needs to be. Even if you know the version number to look for, a search at autodesk.com will come up blank:

Using the Autodesk Knowledge Network Download Finder won’t help, either:

Fortunately, these direct links appear to work:

This brings me to my fourth point of criticism. See that the 64-bit executable has “r2” on it? The one I downloaded on 23 March doesn’t. The 64-bit executables are similar in size to each other but the binary content is different. The 32-bit 2018.0.1 has a date of 22 March and the 64-bit 2018.0.1 r2 has a date of 24 March. So it looks like the patch has been patched, at least for 64-bit users.

Information on this patch-patch is non-existent. Should somebody who downloaded and applied the 64-bit 2018.0.1, download and apply 2018.0.1 r2? Will that work? Do they need to uninstall 2018.0.1 first? How should they do that? Will the 32-bit 2018.0.1 also be updated to r2? Should those users hang off a few days to avoid wasting time or go ahead with what’s there now?

Over to you, Autodesk.

AutoCAD 2018 – why did the DWG format change?

In my review of AutoCAD 2018, I had this to say about AutoCAD 2018’s changed DWG format:

Why does AutoCAD 2018 need a new DWG format? It probably doesn’t. The 2013 DWG format is capable of holding pretty much anything you want… Although Autodesk cites performance reasons with certain drawings, I strongly suspect the new DWG format was introduced purely to make life difficult for competitors, and to encourage wavering customers to stay with Autodesk for fear of losing compatibility. In other words, it seems likely this is an anti-competitive change rather than a technical one.

In a recent blog post, highly respected Swiss-based Autodesk development and research person Kean Walmsley had this to say on that subject:

The main reason for the break in compatibility is some longer-term work that’s going on inside the AutoCAD codebase. For now this is really only surfacing in small ways – I expect it’s contributing some performance benefits, for instance – but the work is absolutely critical to the long-term viability of the product.

Kean’s a straight-shooter and I’m always ready to be corrected if it can be shown that I’m wrong. So I would be interested to learn more detail about this long-term work that’s critical to the long-term viability of the product. It might be good news for customers or really terrible news. If the groundwork is being laid for a file format that’s more heavily cloud-reliant or subject to continuous change, say, that would be an absolute tragedy for customers.

Autodesk is clearly manoeuvring customers into a position of maximum tie-in using various nefarious means, and if the DWG format change is part of that then it’s to be condemned. Maybe further information would help alleviate such concerns. Kean can’t provide that information, and neither can the selected bloggers who were given some insight under NDA last week, but I’m sure someone at Autodesk could. That is, if there really is nothing to worry about.

Kean also had this to say:

AutoCAD continues to be a core part of Autodesk’s business – and it continues to receive significant investment in terms of development resources – but don’t expect that to translate to buckets of shiny new features: AutoCAD’s feature maturity means the investment is rightly being focused in other areas (at least for now).

This had me wondering if Kean mistyped “immaturity”, because almost every AutoCAD feature from the last decade was released immature and only the lucky few eventually got finished. There’s a huge mass of outstanding work left to do in AutoCAD just to bring its existing half-baked features up to scratch, practically all of which could be done without disrupting customers with a new DWG format.

As for the feature set itself being mature, I can’t agree with that, either. Maybe it’s considered mature within Autodesk because of defeatist thinking about what’s possible with DWG-based CAD software? Kean’s comments seem to reinforce that impression. From where I’m standing, the lack of progress in recent AutoCAD releases demonstrates a severe lack of imagination and hunger to improve the product, not any inherent natural plateau in CAD development.

I believe this because Autodesk’s keener competitors have shown that no such plateau exists. Bricsys has proven that it’s very possible to improve an AutoCAD-like DWG-based product out of sight with genuinely useful and productive new features, and they can do it without changing the DWG format. Incidentally, my preliminary tests indicate BricsCAD V17 opens and saves DWG significantly faster than AutoCAD 2018, again without the need for a new format. More on that in a later post.

Back to Kean:

This is a tricky balance – and could easily be interpreted as a big company not caring about (some of) its users and only being interested in milking its cash-cow – but the work happening behind the scenes is significant and I believe will ultimately prove to be of real value to our customers.

Real value? History has taught me to be dubious about that. Many things that Autodesk promotes as being of value to customers turn out to be of net negative value. Time will tell with this one.

Sorry, but I really don’t believe that Autodesk cares about AutoCAD and its users as anything but an income source. I know there are still honest, hardworking, enthusiastic people within Autodesk (like Kean) who want to improve the product on behalf of customers. Good luck to those people, because their efforts are being stymied by management. The results we’re seeing out here in customer land are dismal, and no matter what spin is put on that, it must be disheartening.

Autodesk people, caring about users? Sure. Autodesk, the public listed company, as directed from the top? Nope. Autodesk’s actions and inactions tell me otherwise. Zero cares are given. No words can fix that, no matter who they come from.

AutoCAD really is being treated as a cash cow; hang one of those bells around its neck and be done with it.


(Original image: Daniel Schwen)

AutoCAD 2018 – disable InfoCenter

Here’s another AutoCAD 2018 download, just in case you haven’t had enough.

If you want to disable AutoCAD’s InfoCenter (and you should, unless have a specific reason for keeping it – here’s why), go get the updated download from uber-expert Owen Wengerd’s ManuSoft Freebies page. It works for AutoCAD and AutoCAD LT.

AutoCAD 2018 – there’s already an update

If you downloaded and installed AutoCAD 2018 yesterday and don’t/won’t/can’t have Autodesk desktop app running on your systems, you may already have another download to do, because AutoCAD 2018.0.1 is out.

At the time of writing there is no sign of this update on Autodesk’s main site, but you can get at it using Autodesk Account. Go to Management > AutoCAD > Downloads > Updates & Add-ons.

All that’s downloaded is an executable. No readme, nothing. There is currently no official information about the reasons behind this update, what it includes, what it might affect, how to include it in a deployment, etc. You’ll need to make up your own mind whether to install this update now or wait for information about it. I suggest the latter.

Of course, if you’re using desktop app and allowing automatic updates, you don’t need to worry about any of that that. Just trust Autodesk to not break anything and hope for the best.

What’s the worst that could happen?

Edit: see this post for further information.

AutoCAD 2018 – bear this in mind

Given the dearth of new functionality in AutoCAD in recent years, it’s understandable that Autodesk has taken to claiming credit for the same thing twice. The same features have been touted once for the 2017.1 mid-term update and again as 2018 new features. Even I fell for it, listing linetype gap selection as a 2018 feature in my original review.

For the purposes of reviewing the earlier AutoCAD releases and AutoCAD 2018 as upgrades, I have included the 2017.1 features in 2018, not 2017. Some of those features are praiseworthy, and there have been some minor improvements to some of the 2017.1 features in 2018, but let’s count them just once, please.

Blogs and sites that just regurgitate Autodesk’s take on what’s new in AutoCAD 2018 might inadvertently repeat the double-dipping. Bear that in mind when you read reviews.

AutoCAD 2018 – 2/10, would not rent

It must be March, because another AutoCAD has just been launched. Despite this one being codenamed Omega, I’ve been assured it’s not the final AutoCAD release. The good news is that AutoCAD 2018 is twice the upgrade that 2017 was. But hold your excitement, because the bad news is that 2017 was only a 1/10 upgrade.

Autodesk has continued the well-established tradition of acting like a low-quality sausage machine, reliably popping out consistently unexciting products at regular intervals.

Here’s what’s in this year’s underwhelming sausage:

  • Xrefs now default to relative path attachment and there are a few associated tweaks including find/replace path
  • You can select objects off-screen (handier than it sounds)
  • File dialogs remember settings
  • A couple more dialogs can be resized
  • There are a couple more tiny user interface tweaks
  • The stuff from 2017.1 is included, of course, some seeing minor improvements since then
  • iDrop support is gone (will anybody miss this?)

Thrilled? Me neither. The xref stuff is worthwhile, and Autodesk does deserve credit for not adding any seriously undercooked “headline” features or pointless eye candy. That, along with the 2017.1 stuff, quietly fixing some old bugs, sorting out a few old feature issues and some unsung but worthy API documentation improvements, is just about enough to make this a 3/10 upgrade.

But wait! Autodesk earned itself a minus point to take it from 3 to 2 by inconveniencing customers with yet another API change (so your 3rd party stuff will need updating), and most unfortunate of all, a DWG format change. We had grown used to enjoying a lack of compatibility issues (with plain AutoCAD, anyway – don’t get me started on the verticals), thanks to the 2013 format surviving five releases rather than the traditional three. Now that’s over.

Why does AutoCAD 2018 need a new DWG format? It probably doesn’t. The 2013 DWG format is capable of holding pretty much anything you want, as Bricsys has shown with BricsCAD V17’s ability to compatibly store 3D parametrics and (optional) BIM and sheet metal features within that format. The changes to AutoCAD 2018 don’t come anywhere near that level of complexity. Although Autodesk cites performance reasons with certain drawings, I strongly suspect the new DWG format was introduced purely to make life difficult for competitors, and to encourage wavering customers to stay with Autodesk for fear of losing compatibility. In other words, it seems likely this is an anti-competitive change rather than a technical one.

Oh, and the desktop icon is the same as 2017, so if you run both releases side by side while you’re waiting for your add-ons to be updated, you’ll find it difficult to see at a glance which shortcut is which and what’s already running. Minor irritant, but an unnecessary cheapskate move.

There are alleged to be performance improvements; probably true in places. I did notice a slight improvement in 3D Orbit performance, with AutoCAD 2018 now about up to BricsCAD standards. However, this aspect of AutoCAD 2017 performance wasn’t a problem with the sub-20 MB drawings anyway, so you may not notice a difference.

In my experience with the Beta, AutoCAD 2018 was frequently appallingly slow, particularly start-up time, which was atrocious. Those 14 million lines of code have added up over the years, and AutoCAD 2018 is bloatware; literally ten times the size of the more fully-featured BricsCAD V17 Platinum, which starts up in the blink of an eye. However, I will reserve final judgement on AutoCAD 2018 performance until I have had chance to properly test the production version. I have not factored performance in to my overall mark.

There’s nothing in this release that comes close to justifying the annual maintenance fees, let alone throwing away your perpetual license and signing up for subscription. If Autodesk wants customers to think it will be worthwhile signing up for a lifetime of rental fees in the hope that substantial improvements will result, this weak release represents a singularly unconvincing argument.

Overall, 2/10. This is a sub-mediocre release that shows once again that AutoCAD is in maintenance mode. Autodesk appears completely disinterested in AutoCAD other than as a cash cow, and severely lacking in imagination about ways of seriously improving it. Omega might not be the final AutoCAD release, but given the glacial level of improvement being shown, it might as well be.