| Pages in topic: < [1 2] | Word count in memoQ produces 377 words less than word count in CafeTran!!! Thread poster: Michael Beijer
|
|---|
| Of course (and thank you), but... | Jun 23, 2014 |
For those of us who mostly work for agencies, it will be them who decide the valid wordcount, i.e. the tool used for counting. Our fight should be more on the side of making sure we are paid a decent rate for the wordcounts we get from agencies. We all know how many words we can translate per hour in a certain topic and type of file, so this should be fairly easy to accomplish.
Additionally, the rate is a parameter all agency customers understand and can negotiate on, while the diff... See more For those of us who mostly work for agencies, it will be them who decide the valid wordcount, i.e. the tool used for counting. Our fight should be more on the side of making sure we are paid a decent rate for the wordcounts we get from agencies. We all know how many words we can translate per hour in a certain topic and type of file, so this should be fairly easy to accomplish.
Additionally, the rate is a parameter all agency customers understand and can negotiate on, while the differences in wordcount are difficult to explain and would probably lead to a futile discussion and distrust on the part of the agency customer. As for end customers, all they want is to know what they will pay, and therefore discussing counting methods with them is equally futile.
However, I think all the information shared here is very interesting indeed, so thank you! ▲ Collapse
| | | |
Michael Beijer wrote:
This just goes to show how important it is to be aware of the myriad ways of counting words and fuzzies. So far, it looks like CafeTran and OmegaT produce the highest counts, which is obviously in our favour.
Michael
... if your clients start to use the same tools as you do! I find this quite an interesting discussion because you are doing exactly what you complain your clients do.
I think all modern tools today try to do their best to make it easier for Translators to achieve productivity gains and consistency when they translate. But this is not only to benefit the Translator, this is also to improve the turnaround times for translation work which will benefit your clients as well.
If your client happens to be an Agency, or an organisation that manages their own Translation Memories, and maybe even prepares the analysis themselves too, then the idea is that the analysis also reflects the effort as much as possible by trying to reflect the capabilities of the tools used in the word count. I think this is fair because you are not taking advantage of the translator or the client.
The problems come of course when the tools being used are not the same and therefore cannot take advantage of some of the things that are designed to provide these gains. When this happens I guess you have a few options (in my opinion):
- you either renegotiate the analysis for payment, or
- as Tomás said negotiate a fair rate for that particular client, or
- use the same translation tool as your client, or
- enjoy the reasons you use your choice of CAT tool and live with it...
- ... and maybe this one. If you can't reach agreement, don't take on the work!
The differences between counts is indeed a complicated area and it as much about the simple question of whether "Monday, 23rd June 2014" is one word (as a placeable), two words and two numbers or four words, as it is about the technology you use that makes it possible to count this as one word or four based on your effort.
I imagine you wouldn't complain if all your clients used CafeTran or OmegaT, and then you could go back to memoQ or whatever tool you prefer after these to take advantage of things the other way around.
Regards
Paul
| | | | Michael Beijer 
United Kingdom
Local time: 13:58
Member (2009)
Dutch to English
+ ...
TOPIC STARTER
SDL Support wrote: Michael Beijer wrote:
This just goes to show how important it is to be aware of the myriad ways of counting words and fuzzies. So far, it looks like CafeTran and OmegaT produce the highest counts, which is obviously in our favour.
Michael ... if your clients start to use the same tools as you do! I find this quite an interesting discussion because you are doing exactly what you complain your clients do. I think all modern tools today try to do their best to make it easier for Translators to achieve productivity gains and consistency when they translate. But this is not only to benefit the Translator, this is also to improve the turnaround times for translation work which will benefit your clients as well. If your client happens to be an Agency, or an organisation that manages their own Translation Memories, and maybe even prepares the analysis themselves too, then the idea is that the analysis also reflects the effort as much as possible by trying to reflect the capabilities of the tools used in the word count. I think this is fair because you are not taking advantage of the translator or the client. The problems come of course when the tools being used are not the same and therefore cannot take advantage of some of the things that are designed to provide these gains. When this happens I guess you have a few options (in my opinion):
- you either renegotiate the analysis for payment, or
- as Tomás said negotiate a fair rate for that particular client, or
- use the same translation tool as your client, or
- enjoy the reasons you use your choice of CAT tool and live with it...
- ... and maybe this one. If you can't reach agreement, don't take on the work!
The differences between counts is indeed a complicated area and it as much about the simple question of whether "Monday, 23rd June 2014" is one word (as a placeable), two words and two numbers or four words, as it is about the technology you use that makes it possible to count this as one word or four based on your effort. I imagine you wouldn't complain if all your clients used CafeTran or OmegaT, and then you could go back to memoQ or whatever tool you prefer after these to take advantage of things the other way around. Regards Paul
Paul, to be honest I’m not following you here.
You wrote:
‘I imagine you wouldn't complain if all your clients used CafeTran or OmegaT, and then you could go back to memoQ or whatever tool you prefer after these to take advantage of things the other way around.’
If my client used CafeTran to count the words in the source document I would get paid for more words (than I currently do since they usually count using memoQ or Studio). In what way would I then benefit from (switching back to and) using memoQ to do the translating in?
You wrote:
‘If your client happens to be an Agency, or an organisation that manages their own Translation Memories, and maybe even prepares the analysis themselves too, then the idea is that the analysis also reflects the effort as much as possible by trying to reflect the capabilities of the tools used in the word count. I think this is fair because you are not taking advantage of the translator or the client.’
This is nonsense. Let’s be honest: how many translators know how to use all of their CAT tool’s capabilities? I have been using CafeTran for a year or two now and am still learning new things every day. I’d say the situation is exactly the same for all the other tools: translators buy them and then slowly start learning how to access all of their complex, so-called time-saving features. How many people out there really fully understand Studio and are making optimal use of all of its features? Two? Three? For example, it might take someone a year or more to finally master the art of ‘placeables’ (aka ‘non-translatables’), or of using regular expressions to convert dates or numbers. A year of whose time? Is the translator being paid to learn these things?
Also, a lot of these features look great on paper, but in real life never work 100%, or even 50% of the time, and are in need of constant tweaking and re-thinking. So, no, I don’t think it’s fair to not pay the translator for things like numbers or placeables, just because your tool has a Placeables or Number Conversion tool.
And what’s more (and I have said this before but will just keep repeating it until it sinks in): who pays for the CAT tool in the first place? Yes, I do, the translator. It is supposed to make me money, not the damned agency. It is my tool, not theirs. If it saves anyone any time or makes anyone any money, it is going to be my time and my money. Not that of the agency. But I don’t expect you to understand that because you’re on SDL’s payroll.
Michael
[Edited at 2014-06-23 09:57 GMT]
| | | | Michael Beijer
United Kingdom
Local time: 13:58
Member (2009)
Dutch to English
+ ...
TOPIC STARTER | I agree with you, but ... | Jun 23, 2014 |
Tomás Cano Binder, CT wrote:
For those of us who mostly work for agencies, it will be them who decide the valid wordcount, i.e. the tool used for counting. Our fight should be more on the side of making sure we are paid a decent rate for the wordcounts we get from agencies. We all know how many words we can translate per hour in a certain topic and type of file, so this should be fairly easy to accomplish.
Additionally, the rate is a parameter all agency customers understand and can negotiate on, while the differences in wordcount are difficult to explain and would probably lead to a futile discussion and distrust on the part of the agency customer. As for end customers, all they want is to know what they will pay, and therefore discussing counting methods with them is equally futile.
However, I think all the information shared here is very interesting indeed, so thank you!
Hi Tomás,
You’re welcome! I agree with you, but ... as I see it, our fight should involve both making sure we are paid a decent rate by the agencies and making sure our texts are counted in a fair manner. And a host of other things, for that matter…
For example, who says that it should be the agency that provides the word count? Just because this is what usually happens these days, does this make it right? That’s one of the reasons I’m trying to convince AIT to integrate a fuzzy counter into AnyCount: the freelance translator community needs a standard tool/method of word counting that will benefit us and stand up for our rights. Currently, most agencies use either SDL Studio or memoQ to run their analyses, which I really doubt is a good thing.
I’m trying to raise awareness of this and other topics here on Proz and in the various mailing lists because if none of us translators try to understand the nuts and bolts of these processes and then use this information to represent our interests, do you really think SDL, Across or Kilgray will do so? I don’t. It is not in their interest.
Michael
[Edited at 2014-06-23 10:14 GMT]
| | |
|
|
|
| I disagree.... | Jun 23, 2014 |
... and not because I work for SDL. I imagine I am of course influenced by this, as you are because you don't... but I do try to provide my own opinion and be as honest as I can.
Michael Beijer wrote: Paul, to be honest I’m not following you here. You wrote: ‘I imagine you wouldn't complain if all your clients used CafeTran or OmegaT, and then you could go back to memoQ or whatever tool you prefer after these to take advantage of things the other way around.’ If my client used CafeTran to count the words in the source document I would get paid for more words (than I currently do since they usually count using memoQ or Studio). In what way would I then benefit from (switching back to and) using memoQ to do the translating in?
Well, based on the rest of your post I can see why you might not. The way I look at it is that if I can be paid for a higher wordcount, and use a tool that doesn't require me having to handle every single word manually then I'm better off. But clearly this isn't a valid argument for you. Although I think I'm getting the picture now. You want a tool that treats the wordcount as if everything were entirely manual yet gives you the sort of productivity gains that means you don't have to manually process them. When your client hasn't invested in technology to see the same benefits as you and then provide a fair arrangement both ways I would agree with you, and it's a great way to benefit. But when they have, and if you are willing to take work from them, then a more amicable arrangement should be in place. We're all entitled to make money I think.
Michael Beijer wrote: ‘If your client happens to be an Agency, or an organisation that manages their own Translation Memories, and maybe even prepares the analysis themselves too, then the idea is that the analysis also reflects the effort as much as possible by trying to reflect the capabilities of the tools used in the word count. I think this is fair because you are not taking advantage of the translator or the client.’ This is nonsense. Let’s be honest: how many translators know how to use all of their CAT tool’s capabilities? I have been using CafeTran for a year or two now and am still learning new things every day. I’d say the situation is exactly the same for all the other tools: translators buy them and then slowly start learning how to access all of their complex, so-called time-saving features. How many people out there really fully understand Studio and are making optimal use of all of its features? Two? Three? For example, it might take someone a year or more to finally master the art of ‘placeables’ (aka ‘non-translatables’), or of using regular expressions to convert dates or numbers. A year of whose time? Is the translator being paid to learn these things? Also, a lot of these features look great on paper, but in real life never work 100%, or even 50% of the time, and are in need of constant tweaking and re-thinking. So, no, I don’t think it’s fair to not pay the translator for things like numbers or placeables, just because your tool has a Placeables or Number Conversion tool.
I disagree here too. I'm not talking about AutoSuggest and Muses and Machine Translation and all that stuff. I'm talking about how words are handled. So in Studio if "Monday, 23rd June 2014" is recognised it requires no effort or detailed learning from a translator at all. It underlines the words and you can place a localised date into the target with one keyboard entry. If it's not recognised then it's not counted as one placeable anyway. These types of features are what determines some differences in the wordcount. This is why I said if you don't use the same tools then problems can occur. So if you use a tool that doesn't have this out of the box recognition capability then you will have to either handle it as more words, more keystrokes, or learn how to adapt your chosen tool to take advantage of it too.
I think this is perfectly fair. We're talking about professionals here who owe it to themselves to learn how to make the most of their tools. If they are indifferent to this sort of thing it's a personal choice.
Michael Beijer wrote:
And what’s more (and I have said this before but will just keep repeating it until it sinks in): who pays for the CAT tool in the first place? Yes, I do, the translator. It is supposed to make me money, not the damned agency. It is my tool, not theirs. If it saves anyone any time or makes anyone any money, it is going to be my time and my money. Not that of the agency. But I don’t expect you to understand that because you’re on SDL’s payroll.
Do SDL, or memoQ, or anyone else give their software away for free to everyone apart from translators?
Regards
Paul
| | | | | I'm afraid I don't understand | Jun 23, 2014 |
SDL Support wrote:
So in Studio if "Monday, 23rd June 2014" is recognised it requires no effort or detailed learning from a translator at all.
So if that's not counted, how does it benefit the translator who bought Trados? Does she pay for features that only benefit the agency, or doesn't she?
Cheers,
Hans
| | | | | You're missing my point. | Jun 23, 2014 |
Meta Arkadia wrote: SDL Support wrote:
So in Studio if "Monday, 23rd June 2014" is recognised it requires no effort or detailed learning from a translator at all. So if that's not counted, how does it benefit the translator who bought Trados? Does she pay for features that only benefit the agency, or doesn't she? Cheers, Hans
This is not intended to be a dig at anyone, I just think wordcounts needs to represent the interests of all parties who have invested into technology to improve their ROI. So let me try and explain what I mean here:
Example #1
"Monday, 23rd June 2014" is analysed in a tool that counts these as four words (for example).
Translator handles the work in a different tool that sees this as a single placeable (maybe the one who bought Studio) and is able to handle the words with a single shortcut. Translator gains.
Example #2
"Monday, 23rd June 2014" is analysed in a tool that counts these as a single placeable or as four words.
Translator handles the work in the same tool and all is well. Happy days.
Example #3
"Monday, 23rd June 2014" is analysed in a tool that counts these as a single placeable.
Translator handles the work in a tool that sees these as four words. Translator loses.
So this is not really about translators vs agencies, although it may seem that way if you are always at the end of Example #3. It's about using the right tools for the job when both parties are trying to save money and have invested into technology to help them achieve it.
So in this scenario (translator vs agency) there is no fair solution unless both parties use the same tool, or an independent agreement is made on the wordcount. Maybe something like the GMX standards, although it's quite interesting nobody brought this up and is perhaps indicative of it's current status as a standard for wordcounts, and of the importance people place on this in practice.
I doubt very much that Agencies will be using tools that generate higher wordcounts to bill their clients, and then pay their translators lower rates based on the tools Michael is complaining about. They are more likely going to be negotiating rates because at the end of the day everything comes down to money, which is where Tomás is coming from. As long as you get paid enough to make you happy based on what's in your bank account then the wordcount becomes less of an issue anyway.
Regards
Paul
| | | | | a few answers | Jun 24, 2014 |
Hi,
I would like to point out a few things in this discussion that I would challenge. I'm the guy who always praised memoQ's Parallels compatibility and then wrote to Hans that "we never supported network drives officially". I take complete credit for that, but would like to give the context: we released memoQ 2014 which we never tested on a Mac, as that was not part of our test descriptions. This version broke on network drives which are crucial for Parallels users. Hans wrote a me... See more Hi,
I would like to point out a few things in this discussion that I would challenge. I'm the guy who always praised memoQ's Parallels compatibility and then wrote to Hans that "we never supported network drives officially". I take complete credit for that, but would like to give the context: we released memoQ 2014 which we never tested on a Mac, as that was not part of our test descriptions. This version broke on network drives which are crucial for Parallels users. Hans wrote a message asking what's up, and I answered that we're sorry, we'll fix this, but we never supported network drives officially, so we don't test for that. Two weeks later we released the next build, and now this problem is gone. I think it's really dangerous to place text without context, because this is completely not what I wrote, even though that one sentence is correct.
A few ideas:
- first, word counts. In a lot of regions (Central Europe, in a way Germany, etc.) payment is based on character (in Germany there are standard lines, which is number of characters with spaces divided by 55 or 60 if I'm not wrong). While the translation industry is driven by English source text and as a result British/American payment methods where word is the unit, all the product designers of Kilgray think that actually characters are way better to estimate text. What Paul wrote down with the date, that's completely undisputable with characters. Also, a chemical compound can have 25 dashes in it, just like "back-end" which can also be spelt as "backend" or "back end". Is it a word or two or 25? It is a given amount of characters. Also think about Chinese, Japanese, etc. where word counting is quite a strange concept and Korean where words were only "introduced" (spaces were introduced) recently.
You can read more about this here:
http://kilgray.com/files/user-guide/WordCounts_and_matchValues_EN.pdf (very useful stuff by Angelika Zerfass!)
http://kb.kilgray.com/article/AA-00365/0/Why-are-my-memoQ-word-counts-different-from-Microsoft-Word-or-other-translation-tools.html
- when you are testing word counts or anything, the best method is to understand what is different before going into complaining about one tool or another. The only way to do this is to create a document yourself, with those items that you think may be different. You can take a document, cut that into two halves, and check what are the recurring patterns if the results are more similar. There are many ways to do this. When we were coming up with the "Trados-like" word counts (modeled after Trados 6.5, ancient version), we checked a lot of things and found that some things made sense, but for example why certain standalone symbols like @ were treated as a word whereas others were not did not make sense to us. Studio and memoQ don't have this problem anymore, but certain word count changes may also be related to file filters.
- I completely disagree with our focus on LSPs and large players. We actually have a policy of balanced benefits, every release must contain several things for freelancers, LSPs, and enterprises. memoQ 2014's local project templates, better SDL / STAR package handling, duplicate filtering for TM, TB, improved tag handling, better subsegment capitalization, etc. are pretty much freelance-focused, and I think that translators don't mind the import of embedded images and embedded projects either, even though that's identically important for the other groups. As a matter of fact, memoQ 2013 R2 had a bias towards translators, with the only major organizational feature being package handling from online projects. It is NOT because we have more revenue from translators. It is because 1. translators can recommend tools to their customers, 2. translators also work at companies, not every translator is freelance. Some LSPs are translation agencies, but for many, actual productivity is the key more than the 2-3% savings. On the other hand, you guys here are special: you buy a tool because you find it's helping you more. It would be a good idea to start a poll why translators buy a tool... You would find that maybe you belong to the 20-25% only. Most translators buy a tool because they don't want to learn the ins and outs of importing-exporting files from other tools and troubleshooting if anything goes wrong. I trust that Cafetran's developer is providing excellent support, but in the past, not all tool vendors were famous for this, even among the smaller players.
- what I often find amazing is how much the economic models of fashion work. Today it is much harder for us to keep up the buzz around the tool than to keep up the good quality of memoQ. It's parallel to politics in Hungary: until the left wing was governing the country, I never really made a point of my being liberal in political views. Since the right wing came on power and the left wing is completely ruined, with the right wing having two-thirds in the parliament twice in a row, I am speaking about my "minority views" clear and loud. And yes, the right wing government does a few things right (and many things wrong), but I really don't want to talk about those achievements
Michael, I would be very interested in seeing what's special about the first file. The second was quite close I guess. ▲ Collapse
| | |
|
|
|
Michael Beijer
United Kingdom
Local time: 13:58
Member (2009)
Dutch to English
+ ...
TOPIC STARTER
Sorry about that. I get carried away by my own rhetoric sometimes. I don’t actually believe that you guys are evil (only SDL is truly evil). You are still my second favourite CAT tool, which says something because I have tried many. The problem isn’t really whether or not you guys are aiming your development at ‘us’ (the freelancers) or ‘them’ (the evil corporations), but something else. Let me try to illustrate this with something that happened to me this very morning.
... See more Sorry about that. I get carried away by my own rhetoric sometimes. I don’t actually believe that you guys are evil (only SDL is truly evil). You are still my second favourite CAT tool, which says something because I have tried many. The problem isn’t really whether or not you guys are aiming your development at ‘us’ (the freelancers) or ‘them’ (the evil corporations), but something else. Let me try to illustrate this with something that happened to me this very morning.
A regular(ish) client asked me if I wanted to do a job for them. Well, actually some robotic entity in their stoopid TMS did, but that’s another story. It was nice job, really big, pretty well paid, and in a subject area that I know well. However, it had to be done in memoQ. Hmm, I thought, I’d rather do it in CafeTran, but what the hell, I still own a licence for a relatively recent version of memoQ, maybe I can do it in CafeTran, export the TM, and use that to pre-translate and finalise the job in memoQ. Hell, I might even be able to export the memoQ .xlf and do that in CafeTran, checking and finalising it afterwards in memoQ. However, the file they sent me (a memoQ generated .xlf file*), which they insisted I had to use, was a complete &^%$£ mess. And this is somthing I see all the time. Most of its segments contained at least 5 (very long) sentences and all kinds of tags. I had to shrink down my grid font size in memoQ just to be able to fit a single segment on my screen. Hmm. Simultaneously, I got another job request, from a new client (a freelance translator, like myself). The text looked interesting, also right up my street. Same rate as the job above. I promptly emailed the first client and said ‘Sorry, I am no longer available for job #46747674676474 in your TMS.’
God knows how many times a week I have to turn down large jobs that look great and are pretty well paid but have to be done in either memoQ or Studio. After I politely decline, they then go looking for someone else who will work with their shitty files. However, keep in mind that I was their first choice. They now go and try their second, third, etc. choice(s), all because it simply has to be done in memoQ or Studio. These people are shooting themselves in their own feet. I know quite a lot of mediocre translators who are all too happy to slavishly follow The Rules and who will do the job in whatever they are told to do it in, even if this means they can’t even join or split segments, have no access to their own TMs and TBs, and god knows what other kinds of limitations imposed on them by the tools they are being forced to use. If I close my eyes I can see hordes of ‘Certified Trados users’ who can’t even open a .rar file. Great. The future is bright.
OK, I’ll shut up now, but I hope you see where I am coming from.
If you want I could send you the file in question, the first file (if you want it just PM me!), so you can see what went wrong. I’ve since finished and sent off the job, but it’s just a bit of a bummer getting paid 377 words less for it just because the agency counts everything in Studio/memoQ.
Which brings me to my next point (occasioned by your and especially Paul’s responses). Perhaps it would be a good idea to implement some kind of switch in memoQ/Studios’s counting/analysis tool that PMs could turn on if the job was being counted in Studio/memoQ, but being translated in another tool. Now that would be fair.
Michael
* this is what the file looked like inside (spaces added so it displays properly here):
< ?xml version="1.0" encoding="utf-8"? >
< xliff xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="urn:oasis:names:tc:xliff:document:1.2" version="1.2" >
< file source-language="nl-NL" target-language="en-US" original="ivliddnf" datatype="xml" >
< header >
< tool tool-id="easyling" tool-name="easyling" / >
< / header >
[Edited at 2014-06-24 13:10 GMT] ▲ Collapse
| | | | | That was an Easyling file :) | Jun 24, 2014 |
Hi Michael,
I guess many will envy you for the luxury of being able to turn down a lucrative job Jokes aside, if you look at this, the other party, the LSP wants you to use memoQ just to make sure that they don't have extra work when you deliver in the tool of your choice. Well, probably not you, but our support spent already a Christmas day helping an American LSP who were tolerant enough to give the possibility for ... See more Hi Michael,
I guess many will envy you for the luxury of being able to turn down a lucrative job Jokes aside, if you look at this, the other party, the LSP wants you to use memoQ just to make sure that they don't have extra work when you deliver in the tool of your choice. Well, probably not you, but our support spent already a Christmas day helping an American LSP who were tolerant enough to give the possibility for translators to use Trados bilingual format, and they used some tool that completely messed it up, and in the end it was manual copy-paste. We were not obliged to help, but we went the extra mile, and on a Christmas day there is not too much happening in support anyway...
I checked the XLIFF and this was created by Easyling. If you have trouble with it, please reach out to them, it's not only a very interesting tool but also the people behind it are very approachable and nice. It wasn't memoQ XLIFF but Easyling can create XLIFF for memoQ. I guess it's still the teething problems and nothing else, just talk to them, they'll be happy to listen. It's also a very interesting tool!
István ▲ Collapse
| | | | | Pages in topic: < [1 2] | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Word count in memoQ produces 377 words less than word count in CafeTran!!! | Trados Business Manager Lite | Create customer quotes and invoices from within Trados Studio
Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.
More info » |
| | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news related to CAT tools
