Dominik Kreuzer had some interesting comments
on Software as a Service (SaaS), which includes TEnTs like Lingotek, Elanex,
and XML-Intl, many translation project management solutions (such as Plunet,
Beetext,and Worx), and even online-based tools like
search engines or online dictionaries:
While SaaS may make sense for corporate users, freelance translators and
smaller companies may be severely disadvantaged by the concept on several
fronts. Apart from the speed and customizability issues (...) there's the
question of privacy: With SaaS, it would be easy, for example, to monitor the
time a translator spends on a job and use this info to force down prices. But my biggest concern is the immediate dependency on a whole chain of
providers and hardware: Any fault or problem between PC and SaaS provider's
installation can cause the translator's work to come to an immediate and
complete halt. British Telecom can take a week or more to repair local storm or
water damage to phone lines. No need to spell out what that would mean in terms
of lost income and missed deadlines. If the SaaS publisher ceased operating completely for whatever reason,
the software may be gone forever - complete with the remotely hosted data,
unless the user has taken the precaution of making local backups in a generic
format. No chance for a gradual transition to a new product. But it doesn't take financial or national crisis to stop a translator in
their tracks: if a newly installed service patch causes problems on a local
installation, you just return to the previous release until the bug is fixed or
you've had a chance to find out where the problem lies. With SaaS you are --
again -- dependent on other people to sort it out.
I think that his
points are well taken, but I would still like to add a couple of things.
First, the privacy
issue: While it may be possible to monitor time better, there are a number
of ways to do this even now (think of MS Office's time tracking features
in Properties, or time stamps on comments and tracked changes), and I
doubt that they are used very much. I hope that we are mature enough as an
industry to understand how much time on average (!) it takes to translate,
edit, proofread, format, etc. Plus, I have worked as a project manager myself,
and I gotta tell you, there is no time for the typical PM to track the minutes that
individual translators spend on their assignments.
Interruption of
service: This is a very valid
concern and sadly points to one of the few areas where there is a real
discrepancy between urban and rural areas -- when otherwise it's no problem for
many of us to live in the boonies. I would always advise having a backup
service provider. This is no problem in a city where you have a choice between
many different ISPs, many of whom use different data highways. In rural areas, however,
you may have only one or two providers that often use the same data access
routes, which could mean simultaneous loss of access.
Disappearance of a
SaaS provider: Also a good
concern, and so the first thing you need to ask is in which formats you can
download data backups. When Lingotek first came on the market, it
offered neither TMX nor XLIFF (exchange formats) to download translation data,
which was like an automatic no-go. Many protested, including this newsletter,
and those formats were quickly introduced. Realistically speaking, a certain
feeling of uncertainty about the service could end up making a user a better
data-backer-upper than the user of a PC-based service. With PCs the danger is
less that a program may suddenly not be available, but if your PC goes kaput,
your data may also be in the can.
Messy patching
services: Also a good point
and that is why it's important to "court" your tool provider for some
time to understand whether they have the technical expertise and appreciation necessary
to quickly do a rollback if needed.
All this said, most of
the tools that I use myself are not SaaS tools, but I still believe that this
is where we are headed. And I don't think that this will take 10 or 15 years to
become a mainstream solution. Instead, expect it to be standard in the next
couple of years.
|