Possible issue with upgrading to LR4 - READ

Joined
Feb 5, 2009
Messages
2,390
Location
NC
I've come to the conclusion that updates/upgrades are simply a "how lucky do you feel?". The unlimited number of individual configurations that exist with us today are unreal.

I'm typically a wait and see person, but I downloaded LR 4 today, and no issues thus far.....tomorrow, I may have to revert to backup....it is what it is.
 
Joined
Nov 14, 2007
Messages
1,431
Location
Laurel, MD USA
I didn't have any problems. And I didn't read about a problem until after I did the upgrade. Guess I was one of the lucky group.
 
Joined
May 3, 2007
Messages
6,387
Location
Colorado Springs, Colorado
Pretty easy to protect against. Just save your updated LR3 to LR4 catalog(s) in a separate directory (so they are easily found). If there are issues, simply go back to using LR3 until things are fixed and then simply delete the LR4 catalogs and reopen and update the relevant LR3 catalogs with LR4.

And, it goes without saying, don't do any critical work with LR4 until this is fixed.

I do think that Adobe might want to rethink its policy of disabling catalog updates during beta. This issue would almost certainly have been found and resolved in beta if they had allowed beta testers to use copies of existing catalogs.
 
Joined
Feb 24, 2006
Messages
10,853
Location
Gulu, Uganda
There is good reason to not allow catalog upgrades Doug including that the PV2012 was still being worked on. To allow upgrades would have been irresponsible. It is interesting that there appears to be just a few combinations of products, etc seeing this making it difficult to see during testing. Adobe have a robust prerelease testing procedure however nothing is perfect.
 
Joined
May 3, 2007
Messages
6,387
Location
Colorado Springs, Colorado
There is good reason to not allow catalog upgrades Doug including that the PV2012 was still being worked on. To allow upgrades would have been irresponsible. It is interesting that there appears to be just a few combinations of products, etc seeing this making it difficult to see during testing. Adobe have a robust prerelease testing procedure however nothing is perfect.
Well, I guess we will just have to agree to disagree (in an agreeable Cafe sort of way, of course). :wink:

I have done software testing, run my share of beta tests (albeit "in-house", never with a commercial product). I have also participated in beta testing, both open and "invitation only".

In my view, it would have been entirely within the bounds of responsible testing to have allowed "test upgrades" with appropriately dire warnings and disclaimers. It would have been very straightforward to make a copy/clone of a production environment and then applied the upgrade(s) to that. I suspect that one reason Adobe chose the more conservative course was to insulate their production team from the occasional beta-tester-catastrophe and the associated complaints and noise on the user fora. This is a completely reasonable choice on their part. It simply would have not been my choice had I been running the beta.

Anyway, I am actually quite impressed with the finished product. I find that I am getting good or better results than with LR3 and in fewer steps and less time.
 
Joined
Mar 11, 2008
Messages
1,431
Location
Alaska
Here's my example - http://imgur.com/a/4eIgN

In LR4 after upgrading catalog from LR3.x

Image one is how I edited it in LR3

Image two is after clicking the ! and setting the camera cal from 2010 to 2012. Notice how the numbers stay the same in tone curve, but the medium contrast is reset to linear.

Image three is same as image two, but I switched linear back to medium after the conversion.
 
Joined
Jan 26, 2005
Messages
15,253
Location
Marysville, WA
There is good reason to not allow catalog upgrades Doug including that the PV2012 was still being worked on. To allow upgrades would have been irresponsible. It is interesting that there appears to be just a few combinations of products, etc seeing this making it difficult to see during testing. Adobe have a robust prerelease testing procedure however nothing is perfect.
While I agree with you, Geoff, that in the initial Beta testing this would make sense, but there is no good reason to not include at least a second round of Beta testing that would include such an important piece of functionality. Having been a Dev manager of a product much more expansive, and far more expensive, than Lightroom I can tell you that I would never have allowed this big a miss, I would have been yelling and screaming at the executive level if important functionality was not tested thoroughly.

Unfortunately we are seeing far too much of this with software these days, my guess is that it is a marketing decision to meet a release date vs. an engineering decision.
 
Joined
Nov 17, 2007
Messages
2,192
Location
East of Everett Washington USA
Here's my example - http://imgur.com/a/4eIgN

In LR4 after upgrading catalog from LR3.x

Image one is how I edited it in LR3

Image two is after clicking the ! and setting the camera cal from 2010 to 2012. Notice how the numbers stay the same in tone curve, but the medium contrast is reset to linear.

Image three is same as image two, but I switched linear back to medium after the conversion.
Hmmmm... not good IMO and I'm a LR fanboi.
 
Joined
Jul 21, 2007
Messages
5,262
Location
NJ
While I agree with you, Geoff, that in the initial Beta testing this would make sense, but there is no good reason to not include at least a second round of Beta testing that would include such an important piece of functionality. Having been a Dev manager of a product much more expansive, and far more expensive, than Lightroom I can tell you that I would never have allowed this big a miss, I would have been yelling and screaming at the executive level if important functionality was not tested thoroughly.
At work I was appointed mentor to a new hire. As we're testing a new model he had built I feed it some unusual but not exotic parameters, and the whole model turns into a smoking heap of debris.

"Test failed", the new hire concludes.

"No Chris, this is why we test"

I fully agree with what you wrote and I am absolutely appalled that a blunder like this, as straightforward and easy reproduceable as it is (that part it's really important. This is not some exotic bug that only shows up on a 32bit Windows Vista box with 8 GB of ram, a 15GB hard disk and only on a full moon with outgoing tide).
How this managed to sail through beta testing is beyond me.
 
Joined
May 3, 2007
Messages
6,387
Location
Colorado Springs, Colorado
At work I was appointed mentor to a new hire. As we're testing a new model he had built I feed it some unusual but not exotic parameters, and the whole model turns into a smoking heap of debris.

"Test failed", the new hire concludes.

"No Chris, this is why we test"

I fully agree with what you wrote and I am absolutely appalled that a blunder like this, as straightforward and easy reproduceable as it is (that part it's really important. This is not some exotic bug that only shows up on a 32bit Windows Vista box with 8 GB of ram, a 15GB hard disk and only on a full moon with outgoing tide).
How this managed to sail through beta testing is beyond me.
Well, I disagree a little. I can fully understand how this happened.

All software development is, in essence, a business problem. You have a finite resource pool (money, people, time to market), a target set of features, and a price target. My guess (and it is just a guess) is that, perhaps based on earlier experience, Adobe chose to disallow catalog conversions to protect their development resources from the (likely) complaints and "bug reports" from beta testers who failed to heed warnings about the risks of conversion and admonitions NOT to use the beta for production work. With a public beta it is a statistical certainty that some folks will simply not heed such warnings and that some will report their self-inflicted wounds as bugs.

Further, beta testers tend to be early adopters and, as such, influencers within the user community. If they get cranky and bad mouth the product, there can be economic consequences. Again, I think there are very understandable reasons that Adobe might choose the risk of overlooked deficiencies versus the consequences of user misuse of the beta and data corruption.

As noted earlier, if it had been my decision my bias would have been toward allowing catalog upgrades and accepting the risks inherent in such a decision. However, I freely admit that if I had been privy to all the information that the Adobe project managers had, I might have made the same decision they did. At any rate, Adobe has a pretty good track record of fixing things promptly so this should soon be a non-issue.
 
Joined
Jul 21, 2007
Messages
5,262
Location
NJ
Well, I disagree a little. I can fully understand how this happened.

All software development is, in essence, a business problem. You have a finite resource pool (money, people, time to market), a target set of features, and a price target. My guess (and it is just a guess) is that, perhaps based on earlier experience, Adobe chose to disallow catalog conversions to protect their development resources from the (likely) complaints and "bug reports" from beta testers who failed to heed warnings about the risks of conversion and admonitions NOT to use the beta for production work. With a public beta it is a statistical certainty that some folks will simply not heed such warnings and that some will report their self-inflicted wounds as bugs.
(...)
At any rate, Adobe has a pretty good track record of fixing things promptly so this should soon be a non-issue.
Everyone likes to villify Microsoft. But if you talk to people who work for Microsoft, even lowly sales reps, you'll notice that they're all using their beta's for everything. "Eat your own dog food" it's called. And public beta's are just that - public beta's that only get pushed out after all fatal flaws have been wrinkled out. I remember vaguely when I pointed out that using a beta involved certain risks I was told that that would not be the case. Now I can see why: because Adobe decided not to test "risky" stuff. That's not really testing.

So Adobe decided to ship on a product where a major part of the software - installing it as an upgrade - was not properly tested? Again, this defies belief to me. Why would you do that? Is "the public" the only beta testing ground for Adobe? I'm sure it's going to be fixed quickly but it's rather embarrassing to push out a bug fix so early.

In regards to fixing things quickly: the one thing I'm not using Photoshop for is editing screenshots (when writing documentation). Not because other software (like snagit) is better (although it is, but installing software on a locked down corporate PC is a major disaster so I'd rather not), but because "prt scr and paste" doesn't work - Photoshop doesn't let me paste anything from "outside" photoshop.
Which is a complaint on the adobe support forums since CS3 I think. Really quickly...
 
Joined
Feb 5, 2009
Messages
2,390
Location
NC
At least it is not the disaster that occurred in Apple's Aperture 2 to Aperture 3 upgrade. I waited 3 months after purchasing the upgrade before considering installing it. Once I did, it was an ongoing "good Aperture/bad Aperture" scenario. Lightroom 3 saved the day, and my sanity. :smile:
 
Joined
Jan 26, 2005
Messages
15,253
Location
Marysville, WA
Guess all ya like Bill but you have no inside information on how it works at adobe or how it worked this time.
Frankly, Geoff, my comments are based on Best Practice for any software development organization, and if Adobe chooses to not follow them, I will state:
"Shame on you Adobe".

And I do know "how it worked this time", that is blatantly obvious that not doing any testing of importing a catalog did not surface this bug. It is also obvious that even a small bit of Beta testing would have surfaced this bug. What is not so obvious is why the Lightroom internal QA missed this. My guess, again just my guess as I have no inside knowledge, is that no specific use case and QA test was created which would look specifically at each and every adjustment that could be made. Again, this is against Best Practice in softare development, shame on Adobe. Now, it could be that they ignored this on purpose, it is not uncommon at all that many bugs are not fixed on initial release. Those, however, are the ones that people are unlikely to run into.

You can try to defend them all you want, it does not change the fact of them not following best practice.

You are correct, it is my guess that this is marketing driven, and that guess is based on 40+ years of software development. As Doug correctly notes, this is a business issue, and it is quite obviously Adobe's choice to not bother Beta testing such a crucial piece of code.

As I stated in my first reply, I am seeing far too much of this in software these days, not just Adobe. Just ran into a Microsoft Excel one the other day as a matter of fact. Sad state of affairs, in my opinion.
 
Joined
Jan 26, 2005
Messages
15,253
Location
Marysville, WA
Everyone likes to villify Microsoft. But if you talk to people who work for Microsoft, even lowly sales reps, you'll notice that they're all using their beta's for everything. "Eat your own dog food" it's called....
Having spent about 50-70% of my working hours over the last 7-8 years on-campus at the Microsoft Main campus in Redmond, I can tell you this is very true. The interesting thing is listing to the Microsoft employees vociferously discussing the state of the software in those early releases. If one were to say "Expletives Deleted" :eek: it would be a vast understatement.

Your comments are, to steal a phrase from "across the pond", Spot On.
 
Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Nikon Cafe is a fan site and not associated with Nikon Corporation.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2005-2019 Amin Forums, LLC
Top Bottom