Suggestions on speed complaints

I just downloaded the latest L4 Win beta that has some great optimization in it. Might I suggest that serious focus on optimizing L4 Mac functions before adding additional features may quell the complaints on speed and beach ball issues?

Yes it is functional, but compared the to Win side it lacks speed in Syntax searching, morph searching, parallel resources, parallel passages, commentary section of the PG and word by word in the EG.

I know you are all working very hard and have brought this product a long way. Perhaps the focus from feature to optimization may be in order.

Thanks for all you do!

Find more posts tagged with

Comments

mab

I'm inclined to agree with you John, but I think the programmers necessarily deal with a full plate regardless. Optimization usually follows what works. Optimizing one thing might break something else. I used to write some blisteringly fast code--only to want to find a cave to hide in when I forgot about the rest of the program it had to run in.[:$]

Johnny Sewell

I am having the slowness and beach ball issues more tonight than before with the MacBook Pro 2,2 2.16 GHz.

4.0b Beta 2

Mon 18 Oct 2010

4.10.4.6567

Rosie Perera

https://community.logos.com/discussion/comment/184401#Comment_184401

Optimization usually follows what works. Optimizing one thing might break something else.

That's precisely why "optimizing" shouldn't be a separate step that happens after the code is designed. If you plan the data structures, memory usage, and algorithms well up front, you won't have to do as much post-production optimizing which could break something. Nor will you have to throw out slow code and write it all over from scratch. Just doing minor optimization tweaks afterwards often doesn't gain you enough performance improvement to be noticeable. If you choose to write a heapsort or a quicksort instead of a selection sort, you'll be much better off in the long run. But if you always "code first, optimize later" you might end up wasting a lot of development time, or you might be locked into a poor (slow) design that you can't change because it would "break everything." This is my pet peeve about Logos's development strategy, which I like to give them a friendly nudge about every once in a while. [:)]

mab

https://community.logos.com/discussion/comment/184812#Comment_184812

Not sure how they QA'd L4 in the planning stages. The company I used to work for only took QA seriously in early development after decades of letting QA bring in the mops for their sloppiness. Too many company coders still thought QA was the equivalent of the evening janitorial staff.

John Fidel

https://community.logos.com/discussion/comment/185165#Comment_185165

Actually L4 Win and L4 Mac share much of the same code. L4 Win just went through a significant increase in speed and efficiency.. so it shouldn't be that hard, just a matter how it falls in Logos priorities.

Rosie Perera

https://community.logos.com/discussion/comment/185165#Comment_185165

Not sure how they QA'd L4 in the planning stages. The company I used to work for only took QA seriously in early development after decades of letting QA bring in the mops for their sloppiness. Too many company coders still thought QA was the equivalent of the evening janitorial staff.

Logos still views QA as something to let the beta testers do. They don't have an internal QA team. Developers do their own testing, and there's Melissa. I know Bob has mentioned that he'd like to hire someone to be responsible for internal testing, but he's had his hands full filling all kinds of other positions that were needed -- more customer service reps to handle the upswing in support calls they started having after L4 shipped, more programmers, etc. It still saddens me that QA is a lower priority than developing new products, putting Logos on more platforms, adding more features, and revamping the website, much as all of those things are worthy investments of developer time.

They don't even make a serious attempt to eradicate most of the bug database backlog (1400+ bugs in it at last count) before releasing each new version of the engine. The list of "known issues" included in the release notes each time makes it seem like they are shipping a near perfect release and only a couple of things are not dealt with yet, when in reality there are a large number of known bugs that they are intentionally putting off. At Microsoft, every time we shipped a new release, we scheduled a ZBR ("zero bug release") internally before going gold. We'd only freeze code changes after we'd bounced down to zero bugs every day for a week or so with no new major show-stopping bugs coming in. And then we'd go gold only after another week of testing revealed no serious bugs. Yes, after a certain point in the development cycle we'd start postponing bugs until the next release if they were really minor, or otherwise we'd be fixing bugs forever and never ship the thing. But anything that was a serious malfunction or would be a major nuisance that users were very likely to run into would get fixed. That is, until we got into the last couple of weeks, when fixing anything but crashing bugs was deemed to be too risky, as it might destabilize the product.

I am really hoping they're serious about 4.2 being primarily a bugfix release. Todd White said we shouldn't expect many new features in it because, among other things: "We want to address some of the 1400+ bugs and 800+ minor features in our bug tracking database." How about addressing 1200 of the 1400 bugs? That would be a good start. We can live with the remaining 200 until 4.3. [:)]

mab

https://community.logos.com/discussion/comment/185190#Comment_185190

QA at beta is the precise problem. QA people need to be an integral part of the design and development team process--something that didn't become clear until maybe the last decade or so. It drives down development time and reduces support costs. Bugs are still inevitable (I "broke" one of my company's major released programs in less than an hour on my first stint as a QA intern). Users will always uncover some of them but they should be fewer. Since Logos is rather complex and getting more so, I hope they allocate for expert QA staff soon.

Bob Pritchett

https://community.logos.com/discussion/comment/185190#Comment_185190

They don't even make a serious attempt to eradicate most of the bug database backlog (1400+ bugs in it at last count) before releasing each new version of the engine. The list of "known issues" included in the release notes each time makes it seem like they are shipping a near perfect release and only a couple of things are not dealt with yet, when in reality there are a large number of known bugs that they are intentionally putting off. At Microsoft, every time we shipped a new release, we scheduled a ZBR ("zero bug release") internally before going gold. We'd only freeze code changes after we'd bounced down to zero bugs every day for a week or so with no new major show-stopping bugs coming in.

Hey, that's not fair. Remember, I was at Microsoft, too. :-)

No product I saw there ever had "zero" bugs -- it had "zero bugs we intend to fix." They didn't delete the others, and the counts on some products (Windows!) went into the tens of thousands, as I recall. So the fact that we've got bug numbers up to 1,400 doesn't mean we're ignoring 1,400 important "we intend to fix" bugs. It means we've logged 1,400 bugs / ideas / features / etc. We do have a process for triaging the bug list before shipping, deciding what we will adn won't hold the product for, etc.

It's not as mature a Q&A process as some companies have, I'll grant you, but it's not quite as sloppy as you're making it sound...

Rosie Perera

https://community.logos.com/discussion/comment/185434#Comment_185434

Hey, that's not fair. Remember, I was at Microsoft, too. :-)

Yeah, I know. :-) But you were a program manager, not a developer. So you didn't necessarily learn the same discipline we did. ;-)

And yes, you're right we did ship lots of bugs. Intentionally. However, as I recall on the Word dev team, we would only not fix a bug if it was not repro i.e., we seriously couldn't reproduce it and even the tester who found could not repro it anymore on their machine; or if it was by design (since most of the testing was done in house for us, "by design" had a much stronger feel to it there than it does for Logos bugs that are resolved that way; for us it didn't mean "thank you very much, but we've decided we not to take your suggestion for how to improve the product " (often Logos users will report feature inadequacies as "bugs" when they really aren't) -- rather, it meant the design specs hadn't been communicated clearly enough and the tester really did think the program was behaving incorrectly but he or she was wrong; or if it was postponed to the next version (we'd only start doing this once we got close to releasing which on our lengthy release cycles might have meant 4-8 months away). We'd only mark a bug as won't fix if it was so obscure that nobody would ever encounter it in a real life scenario or its effect on users was so minor that nobody would ever complain about it except a twisted tester, and it would have been really, really hard and time consuming to fix.

Early on in the dev cycle, any bug was fair game for a developer to work on. They all got assigned to some developer to take a look at (or we could grab one off the bug database and assign it to ourselves if we saw one in our area), and if that developer felt it was easy enough to knock off in his spare time, he just would, period, regardless of any overarching decision of what areas should get the highest priority. Oh, also the tester who reported the bug would assign a priority to it (which could be changed by dev if necessary) indicating how seriously it was interfering with his or a potential user's work. Low priority bugs that were hard to accidentally stumble upon and would be hard to fix were sometimes blown off even in the early stages.

It's not as mature a Q&A process as some companies have, I'll grant you, but it's not quite as sloppy as you're making it sound...

OK, fair enough. But there are bugs I and others reported on the bug wiki page several versions ago which been "confirmed" as bugs and still not fixed. And that's only a fraction of the items in your bug tracking database. I stopped using that method of reporting bugs since it had such a low ratio of actually seeing bugs I'd reported fixed within the next version or two. At MS, we'd also spend some time fixing bugs postponed from the previous release (due to it being too close to shipping time) before adding any new features in a new release. Otherwise they'd just keep on getting postponed indefinitely. In fact, we were constantly working on chipping away at the bug list, not just during bug fixing time towards the end of a release. Our goal was to keep our individual lists of bugs assigned to us down to a screen's worth.

And you know I'm only giving you this hard a time about it because we're friends and I know you know that, and I really care about Logos. (Just in case others watching this exchange think I'm being a jerk.)

Stuart Robertson

https://community.logos.com/discussion/comment/185461#Comment_185461

Are combo-boxes that don't work (http://community.logos.com/forums/t/25003.aspx) in the Mac version part of the 1400 bugs that might be fixed eventually?

I ask (and I know my tone is, perhaps, less than cordial at present) because my Logos for Mac experience has been slowly (but surely) souring my entire Logos experience. I've tried to use Logos for Mac daily for the past week and it hasn't been a great experience - RAM usage rapidly climbs upwards of 900MB (currently sitting at 1.1GB RAM used by Logos for Mac), etc. Many many people post the various defects they find but I get the feeling that Logos users are becoming a free QA and test service for Logos...

When I run Logos for Windows I find that my Logos for Mac workspace isn't synced at times, layouts don't work, etc. Making moving back to Windows inside Parallels Desktop painful too... what is needed is more focus on quality and usability (including performance) and less new features (imho)

Jack Caviness

https://community.logos.com/discussion/comment/185488#Comment_185488

my Logos for Mac experience has been slowly (but surely) souring my entire Logos experience. I've tried to use Logos for Mac daily for the past week and it hasn't been a great experience

I concur.

Cameron Watters (Faithlife)

Yes it is functional, but compared the to Win side it lacks speed in Syntax searching, morph searching, parallel resources, parallel passages, commentary section of the PG and word by word in the EG.

Funny you should mention this. We've actually been spending quite a bit of energy on improving speed in most of the sections you describe. Obviously, different sections will see different amounts of benefit from the various optimizations.

As an example, changes we've been working on in the past few days have reduced the time to generate the WordByWord section of an Exegetical Guide on John 15:1-3 from 1 min 50 seconds on my dev machine to less than 30 seconds.

John Fidel

https://community.logos.com/discussion/comment/185537#Comment_185537

woo hoo. I am looking forward to the next beta. Thanks again for all you are doing to make this a great product.

Rosie Perera

https://community.logos.com/discussion/comment/185537#Comment_185537

As an example, changes we've been working on in the past few days have reduced the time to generate the WordByWord section of an Exegetical Guide on John 15:1-3 from 1 min 50 seconds on my dev machine to less than 30 seconds.

That's great to hear! Is this work being done in shared code that will benefit Win users also? I hope so! [:)]

David Mitchell

https://community.logos.com/discussion/comment/185551#Comment_185551

As an example, changes we've been working on in the past few days have reduced the time to generate the WordByWord section of an Exegetical Guide on John 15:1-3 from 1 min 50 seconds on my dev machine to less than 30 seconds.

That's great to hear! Is this work being done in shared code that will benefit Win users also? I hope so!

Right now, it's all in the (unshared) rendering code. Perhaps it's time for you to get a Mac, too?

Rosie Perera

https://community.logos.com/discussion/comment/185569#Comment_185569

Right now, it's all in the (unshared) rendering code. Perhaps it's time for you to get a Mac, too?

Nope. I'm just going to get a faster PC. [:)]

Stuart Robertson

https://community.logos.com/discussion/comment/185537#Comment_185537

We've actually been spending quite a bit of energy on improving speed in most of the sections you describe. Obviously, different sections will see different amounts of benefit from the various optimizations

Apart from speed can something be done about Logos for Mac's memory usage - dare I say memory leaks. I can use Logos for Windows heavily and its memory usage stays around 450 or 500 MB. With not too much usage and not many windows open Logos for Mac's memory usage climbs steadily until - seemingly without limit. I currently have it sitting at 1.19GB. As the memory usage climbs, speed begins to stuggle.

John Fidel

https://community.logos.com/discussion/comment/185571#Comment_185571

Rosie, you walked into that one![8-|]

Rosie Perera

https://community.logos.com/discussion/comment/185584#Comment_185584

Rosie, you walked into that one!

Yup, I should know better than to go posting questions on the Mac forum. Where's my dunce cap smiley when I need it?

Bob Pritchett

https://community.logos.com/discussion/comment/185461#Comment_185461

Yeah, I know. :-) But you were a program manager, not a developer. So you didn't necessarily learn the same discipline we did. ;-)

Actually, I was a dev before I was a PM. My first internship was on the Windows 3.x team, where I spend 5 months optimizing the HP LaserJet driver. :-)

Alan Macgregor

https://community.logos.com/discussion/comment/185598#Comment_185598

My first internship was on the Windows 3.x team, where I spend 5 months optimizing the HP LaserJet driver.

Well, belated thanks, Bob. I ran Windows 3.1 on my very first PC with my HP laserJet IIIP and never had a moments trouble. Now I know it was all down to you. Thanks again! [:D]

BTW I have(presently) no issues with Logos 4 Mac. It does the business for me in weekly sermon preparation. Though I still use Libronix for academic work (mainly becuase my academic laptop can't run Logos 4).

Every blessing

Alan

Jac Colon

https://community.logos.com/discussion/comment/185579#Comment_185579

Apart from speed can something be done about Logos for Mac's memory usage - dare I say memory leaks. I can use Logos for Windows heavily and its memory usage stays around 450 or 500 MB. With not too much usage and not many windows open Logos for Mac's memory usage climbs steadily until - seemingly without limit. I currently have it sitting at 1.19GB. As the memory usage climbs, speed begins to stuggle.

I am having the same problem. I got up to 1.3GB of memory usage, speed problems and finally after sitting overnight, when I entered a text this morning it crashed on me. I know that it is not a computer problem, I have an 8 Core 2.26 Mac Pro with 16 GB of memory.

I love using Logos 4 but right now, it is getting a little frustrating. I am using 4.0a SR-3 (4.10.4.6316) and am wondering if switching to the most recent Beta would help.

jac

David Mitchell

https://community.logos.com/discussion/comment/185858#Comment_185858

I am using 4.0a SR-3 (4.10.4.6316) and am wondering if switching to the most recent Beta would help.

Switching to the beta because you want more stability or performance is not a good idea.

Graham Owen

https://community.logos.com/discussion/comment/185569#Comment_185569

Right now, it's all in the (unshared) rendering code. Perhaps it's time for you to get a Mac, too?

Personally I have not been a committed user of Logos for all of these years to suddenly buy a Mac one of my reasons for originally choosing Logos was the fact it was a proper PC product. I'm not anti Mac but I made my platform choice some years ago and historically I have had a better user experience using software developed specifically for my preferred platform I really hope that Logos can develop a cross platform product that will work well for both sets of users.

Mike Binks

https://community.logos.com/discussion/comment/186133#Comment_186133

Right now, it's all in the (unshared) rendering code. Perhaps it's time for you to get a Mac, too?

Personally I have not been a committed user of Logos for all of these years to suddenly buy a Mac one of my reasons for originally choosing Logos was the fact it was a proper PC product. I'm not anti Mac but I made my platform choice some years ago and historically I have had a better user experience using software developed specifically for my preferred platform I really hope that Logos can develop a cross platform product that will work well for both sets of users.

Is that Rosie's pseudonym?

Or is he her bodyguard?

[:O]

Rosie Perera

https://community.logos.com/discussion/comment/186208#Comment_186208

...one of my reasons for originally choosing Logos was the fact it was a proper PC product.

Is that Rosie's pseudonym?

Heh, heh. No, couldn't be me. That was never a reason why I started using Logos. The main and in fact only reason I did was because I knew this guy in my division at Microsoft who was writing a little Bible program in his spare time and needed a beta tester. That was Bob Pritchett. The rest is history...

fgh

https://community.logos.com/discussion/comment/186305#Comment_186305

The main and in fact only reason I did was because I knew this guy in my division at Microsoft who was writing a little Bible program in his spare time and needed a beta tester.

You really have been there from the start!

Rosie Perera

https://community.logos.com/discussion/comment/186313#Comment_186313

The main and in fact only reason I did was because I knew this guy in my division at Microsoft who was writing a little Bible program in his spare time and needed a beta tester.

You really have been there from the start!

Yup! [:)]

Eric Smith

My copy of Logos for mac (4.0a SR-3) seems to work well but I did have a problem with it today. several days ago I had changes my customization so that all of the boxes were checked. When I started to have problems with the program I went back to my earlier settings of having only one box checked in customize on the home screen. After that every thing was fast again.

Jac Colon

https://community.logos.com/discussion/comment/186405#Comment_186405

My copy of Logos for mac (4.0a SR-3) seems to work well but I did have a problem with it today. several days ago I had changes my customization so that all of the boxes were checked. When I started to have problems with the program I went back to my earlier settings of having only one box checked in customize on the home screen. After that every thing was fast again.

Fantastic! I did not think it would help me since I never use the home page, but I gave it a try and just as you said, I am back up to speed!!! No more beach balls. Amazing! Thanks so much for sharing.

jac

Jac Colon

https://community.logos.com/discussion/comment/186405#Comment_186405

I just got SR-5 and went back to check the home page boxes again. Everything worked fine. No more spinning beachballs.

At least not yet. Now it is finally beginning to behave like one would expect it to.

Thank you Logos!

jac

Brian Vallotton

https://community.logos.com/discussion/comment/185461#Comment_185461

I have really enjoyed the "friendly" exchange here between you two. Iron sharpening iron comes to mind. [H]

Quick Links

Recent Posts
Activity
Unanswered