Photoshop CC and Lightroom CC/6.2 Updates

[davidc]

Someone made the same comment as you did about how releasing software with known bugs is astonishing - it's actually not, and happens way more than you might expect. I've cherry picked my most relevant thoughts from the same discussion

To be fair to them, and having worked in the software dev industry for many years, I completely understand the reasoning & thinking that will have gone behind the decision to release this update. There's a difference between shipping a bug where you can't reproduce it reliably and you suspect the impact is limited vs. just pushing out a known showstopper issue. It was a judgement call which in this case was the wrong call, and we all like to focus on the negative while we forget the much larger number of smooth updates where there are few tangible issues. It's a mistake, pure and simple, and hopefully they can learn from it.

I think anyone who upgrades immediately without waiting a little while to see if there are issues is also a bit naive and Abobe are far from the only company who have problems with new releases. Apple, Microsoft, Google... everyone does it, we just don't remember the times where it works perfectly.

[Mike Farley]

Like you, I have extensive experience of software development and not just relatively simple applications such as Lightroom, but large complex mission critical systems comprising many components which all have to work together seamlessy. I have observed a number of significant IT failures and had a major role in resolving the aftermath in quite a few of them. Sometimes the causes have been trivial, out of all proportion to the impact, and on other occasions it has been a more complex chain of events which no one could have predicted or tested for. As you say, sometimes problems happen and every company with a significant IT involvement has experienced some form of disruption.

So am I being unfair to Adobe? If this had been a one-off occurrence, I would concur; but for Adobe it is far from unique and there have been a number of dotdotone releases where the company has had to rush out an emergency fix. Four days elapsed between the 6.2 and 6.2.1 updates, so it does not seem that the underlying problem was that difficult to reproduce once there was an incentive. Maybe the premature release this time was simply an error of judgement, most probably driven by demands from marketing, but company shows no signs that it is learning from its past mistakes. It doesn't have to be like this if there is a will to ensure quality (and reputation) are paramount.

[davidc]

Testing is always undervalued, underfunded and squeezed. Bloody developers/project managers

[davidc]

If this is true then they have a lot more work on their hands and the past Lightroom performance issues still haven't gone away!

http://petapixel.com/2015/10/13/lightro ... mpetition/

I knew I should have just stayed with bridge & photoshop

[Mike Farley]

If this is true then they have a lot more work on their hands and the past Lightroom performance issues still haven't gone away!

http://petapixel.com/2015/10/13/lightro ... mpetition/

I knew I should have just stayed with bridge & photoshop

It most likely is true, but it is a selective test and not representative of Lightroom's performance overall. In the main, LR's import process is reasonably fast, although I have heard that it is quicker to copy files from memory cards onto the hard drive manually, then Add them into the catalogue. For some reason LR struggles with images from Fuji's X-Trans sensor and it takes a long time to build the previews, which is where the time is mainly taken. I experimented with putting a faster card into my X-Pro 1, but it did not make much difference as the problem does not seem to be with the transfer speed from the card.

For me, it is not that much of an issue. I usually only import once, twice if I am away and look at the images on my laptop as well, and I can do other things while the previews are being processed. I have not tested it, but there does not seem to be any difference in speed subsequently compared to other cameras once the Raw files have been imported. That said, I am starting to find the X-Pro 1 and 18-55 kit lens to be a useful combo and more of my work is being done with the Fuji. LR works well enough elsewhere that I'll live with the inconvenience of the slow import.

[Mike Farley]

Testing is always undervalued, underfunded and squeezed. Bloody developers/project managers

Testing comes at the end of the development process, unsurprisingly, and unanticipated problems mean that fixed time and budget have already been expended by the time the testers get their hands on the code. They do not improve matters by inconveniently finding bugs which cost more time and money to fix. You are your own worst enemies.

[Mike Farley]

According to Thom Hogan, the bugs which cause the crashes and performances problems are not Adobe's main problem. No, it's the "improvements" to the Import process.

http://www.dslrbodies.com/accessories/s ... pdate.html

For the time being I am going to stick on LR 6.1, apart from the laptop where I am already on 6.2, and see what happens in 6.3. Ironically, because my laptop runs OS X where external media has to be ejected before it is removed to avoid corrupting the card, the now missing auto-eject feature after import would have been very useful. Just what was Adobe thinking?

[Rose]

I too am sticking with 6.1 for the moment... Very grateful that I didn't upgrade for once

[Mike Farley]

I too am sticking with 6.1 for the moment... Very grateful that I didn't upgrade for once

There's only one rash person around here, it seems.

[Peter Boughton]

Testing comes at the end of the development process

And that's a big part of the problem. Testing done properly occurs as part of the development process, and involves more than just humans. When automated tests for each bit of functionality are created before (or alongside) the associated functionality then the number of bugs drops drastically, (and the humans can focus on what is needed to polish the software, instead of relatively trivial issues that automatons can highlight).

I agree with Thom's article - whilst any new key feature/change should be tested more thoroughly, the performance issue is not why there is such an uproar. The hiding and removal of features is the problem - catering solely to the lowest common denominator, rather than playing to the strength of software to provide uniquely tailored solutions.