Are you aware of any further developments regarding the reliability of the TM+ format? I ask because I recently submitted a support case where running stats on a certain local project yielded a number of "Attempt to retrieve deleted entry" errors (a type of error I've never come across before); memoQ support were able to replicate the error and told me the solution is to convert the TMs attached to the project to TM+ format. (Actually they said to wait for the next 11.1.x point release because the current build [11.1.15] has a "known issue regarding TM+").
Obviously I want to eliminate the problem that's generating the error message but I'm reluctant to embrace TM+ if in doing so I'm likely to be inviting other problems…
I have run into a few issues since I published and a memoQ exec birthed a cow over the article, angrily claiming that there were no issues. And I am not at all surprised to hear that 11.1.x has "issues" with TM+. If there are no issues with it in memoQ 12, I'll buy the beer, but I expect to be the one collecting on any wagers since the company now seems fully distracted pursuing AI foolery.
Thanks for responding. After reading your reply, I decided to have a go at running "Repair resource" on the TM to which the error related. Lo and behold, problem solved: no more stats errors. It's interesting (and somewhat annoying) that support leapt straight to recommending conversion to TM+ rather than suggesting a TM repair as an alternative…
It's pure politics. The strongly negative response to my article made it clear that this is not a neutral matter, and some people are pushing an agenda that might not track accurately with the true state of things at present. Otherwise, why would anyone get so heated over it? I'm glad you found a solution in any case. A repair is usually one of the first things I try when a TM gets funky.
TBH I'm not sure why I didn't just try a repair in the first place: on reflection, it's probably because I don't think I've ever previously had a TM issue that required a repair. (When you consider that I've been using memoQ on a near-daily basis since 2008, that's not a bad record!)
Beats my record, but since 2009 I've probably run no more than half a dozen repairs. Considering how much risky stuff I do for testing, that's pretty good.
memoQ is a Windows-based tool. It has never been supported on macOS. Using it on a mac, especially storing anything outside the virtual Windows environment is definitely at your own risk. If it used to work in previous versions, that's pure luck.
The problem here isn't that memoQ was used on a VM on a Mac. The problem is functional incompatibility between TM+ and many ordinary memoQ features that integrate with the TM functionality. I see issues like that frequently, but these days I have neither the time nor the energy to report them as often as I used to. I figure the company has grown enough that they have sufficient resources to do a better internal job of testing such matters.
TM+ is essentially something to serve the bulk market bog and big corporate players using machines with 90+ cores to judge by a remark passed on to me by an attendee of memoQ Fest 2024. It most certainly offers no advantages to people working with TMs of the size I have from my practical work, running to about 400k TUs max. Instead, in its current unfinished state, it creates a host of problems for ordinary users and normal LSPs, and this unchangeable default setting is an unnecessary burden on users at this stage, essentially a forced beta test, because TM+ is really not ready for the normies in the real world. We aren't all out there dropping acid and imagining that training LLMs on insane quantities of data is somehow the future.
care to list a few such incompatibility issues? during the year that has passed since memoQfest 2023, the vast majority of the TM+ bugs has been fixed.
One is discussed in the project TM post and was demonstrated in today's webinar. Further data for that test case will be posted on Thursday. I will also look further into the case that inspired this post and discuss that. I have several more scenarios where I need to check the status of bug fixes and will post those as I go through them.
Hi Kevin,
Are you aware of any further developments regarding the reliability of the TM+ format? I ask because I recently submitted a support case where running stats on a certain local project yielded a number of "Attempt to retrieve deleted entry" errors (a type of error I've never come across before); memoQ support were able to replicate the error and told me the solution is to convert the TMs attached to the project to TM+ format. (Actually they said to wait for the next 11.1.x point release because the current build [11.1.15] has a "known issue regarding TM+").
Obviously I want to eliminate the problem that's generating the error message but I'm reluctant to embrace TM+ if in doing so I'm likely to be inviting other problems…
I have run into a few issues since I published and a memoQ exec birthed a cow over the article, angrily claiming that there were no issues. And I am not at all surprised to hear that 11.1.x has "issues" with TM+. If there are no issues with it in memoQ 12, I'll buy the beer, but I expect to be the one collecting on any wagers since the company now seems fully distracted pursuing AI foolery.
Thanks for responding. After reading your reply, I decided to have a go at running "Repair resource" on the TM to which the error related. Lo and behold, problem solved: no more stats errors. It's interesting (and somewhat annoying) that support leapt straight to recommending conversion to TM+ rather than suggesting a TM repair as an alternative…
It's pure politics. The strongly negative response to my article made it clear that this is not a neutral matter, and some people are pushing an agenda that might not track accurately with the true state of things at present. Otherwise, why would anyone get so heated over it? I'm glad you found a solution in any case. A repair is usually one of the first things I try when a TM gets funky.
TBH I'm not sure why I didn't just try a repair in the first place: on reflection, it's probably because I don't think I've ever previously had a TM issue that required a repair. (When you consider that I've been using memoQ on a near-daily basis since 2008, that's not a bad record!)
Beats my record, but since 2009 I've probably run no more than half a dozen repairs. Considering how much risky stuff I do for testing, that's pretty good.
Anyway, thanks for your help. Good to have an objective view from somebody else!
memoQ is a Windows-based tool. It has never been supported on macOS. Using it on a mac, especially storing anything outside the virtual Windows environment is definitely at your own risk. If it used to work in previous versions, that's pure luck.
The problem here isn't that memoQ was used on a VM on a Mac. The problem is functional incompatibility between TM+ and many ordinary memoQ features that integrate with the TM functionality. I see issues like that frequently, but these days I have neither the time nor the energy to report them as often as I used to. I figure the company has grown enough that they have sufficient resources to do a better internal job of testing such matters.
TM+ is essentially something to serve the bulk market bog and big corporate players using machines with 90+ cores to judge by a remark passed on to me by an attendee of memoQ Fest 2024. It most certainly offers no advantages to people working with TMs of the size I have from my practical work, running to about 400k TUs max. Instead, in its current unfinished state, it creates a host of problems for ordinary users and normal LSPs, and this unchangeable default setting is an unnecessary burden on users at this stage, essentially a forced beta test, because TM+ is really not ready for the normies in the real world. We aren't all out there dropping acid and imagining that training LLMs on insane quantities of data is somehow the future.
care to list a few such incompatibility issues? during the year that has passed since memoQfest 2023, the vast majority of the TM+ bugs has been fixed.
One is discussed in the project TM post and was demonstrated in today's webinar. Further data for that test case will be posted on Thursday. I will also look further into the case that inspired this post and discuss that. I have several more scenarios where I need to check the status of bug fixes and will post those as I go through them.