Hacker News Reader

Comment by Spearchucker | Hacker News Reader

These kind of articles are always a little lop-sided. They state, for example, that "The migration will save the government some 1.5 million euro per year on proprietary software licences."

License savings. Great. What was the cost of:

- Re-imaging those 120,000 desktops?

- Re-training the (I assume) 100,000+ users?

And what are the support costs?

What were they under MS Office? It's laudable, for sure, but I don't believe that F/OSS is cheaper. I'd like to see a real cost impact analysis that covers the entire program, from inception through to on-going operational management. And something that financially compares the as-was state with the current state.

View on HN

I have experience with the Valencian government, specially with the health system department, and in 2008 they were using a 3rd party tool for that (deployment, remote support, monitoring, etc).

AFAIK they've been looking at OSS for years (I was involved in a project related to that), and the main blocker was that the BOFH tool they were using didn't support OpenOffice back then (that was before LibreOffice existed).

In fact moving from MS Office to OpenOffice was the easy part, there was a more ambitious project to migrate desktops to Linux, but the provider of the management tool was only supporting Windows.

> License savings. Great.

That and other related admin such as keeping all installations in check to make sure you are within your license terms. Have you budgeted properly for the time of experienced people to fill out paper work and the potential disruption of everyone else if a full audit is requested (the license agreement says that can call one of those when-ever they like IIRC).

Sometimes using Office most definitely is the right answer (the "compatibility with our suppliers/clients" issue often can't not be ignored for a start) but when it isn't the correct answer pure licensing costs are not always the only reason.

> What was the cost of: Re-imaging those 120,000 desktops?

The costs of this change per desktop will be similar to an upgrade of Office.

If they are upgrading from an older version of office that is no longer supported and the choice is between newer office and OO/LO then this is moot. There may be a bit of extra time at the start of the project as those running it get familiar with OO/LO's installation foibles, but considered per desktop this is not going to be high.

In fact installing OO as a new package may be more risk free as it can co-exist with their current version of Office, so they have an easy fallback/rollback or slow transition option.

> Re-training the (I assume) 100,000+ users?

The differences are not that significant. If they are moving from a relatively old version of office then the retraining for the newest Office variant could easily be more significant due to the UI changes over recent years.

If they have any custom training materials then there will be a cost for updating these, but then there will be for Office too.

> And what are the support costs?

Again, I doubt these will be significantly different. There might be a little more effort needed here due to people being familiar with Office due to the copy they run at home or used at a previous employer, but again there is the "which office version" difference which may make this a moot point. If people using something at home is one of your considerations and you are relying on it to reduce your support costs then OO/LO makes sense as everyone can have an install of OO/LO at home legally for no cost.

>> What was the cost of: Re-imaging those 120,000 desktops?

>The costs of this change per desktop will be similar to an upgrade of Office.

Not necessarily. Depending on the version of Office you're upgrading from, you can do regression tests on a model office machine, and then just pump the upgrade out using WSUS [1].

Good points though.

[1] http://technet.microsoft.com/en-us/windowsserver/bb332157.as...

The re-imaging and re-training likely would've been incurred by an upgrade to a new version of office, so those may not actually be differentiators.

In my experience, the main cost to adopting F/OSS for office software is interoperability with people sending word and excel files, not support or training.

Note, LibreOffice is available in the native language of the users, Valencian, and has spell-check dictionaries for that language. Presumably, the Valencian government puts out most communication in Valencian, so there may actually be savings in not having to develop and maintain their own Valencian dictionary for Office, which Microsoft does not provide, and reduced support as the interface uses terms more familiar to the native Valencian speakers.

>>>Note, LibreOffice is available in the native language of the users, Valencian, and has spell-check dictionaries for that language. Presumably, the Valencian government puts out most communication in Valencian.

TIL people in Valencia speak several different versions of Spanish: http://gospain.about.com/od/learningspanishinspain/p/valenci...

"In Valencia they speak a form of Catalan (the locals call it 'Valenciano' but this is not officially recognised anywhere outside of Valencia).

Everyone in Valencia speaks Castillian Spanish, but many will speak to each other in Catalan in the street. This will be a bit of a hindrance to your learning experience."

"You will mainly hear Catalan in the streets of Valencia, not Castillian Spanish. Even road signs and advertisements are in Catalan. You will find newspapers, TV and radio in both languages."

This website is very wrong. Firstly, there are no such thing as "different versions of Spanish" in Spain. People from different regions speak with (sometimes very) different accents, but it is the same language. Everyone can understand each other while speaking spanish.

"Valencià" (Valencian) is the catalan dialect spoken in Valencia. Catalan itself is a whole other language than spanish. They are as different as, say, spanish and portuguese or italian and french (and any combination of those).

Secondly, Valencian knowledge and use is being lost in the whole region [1]. By 2008, only 32% of the population claimed to use it as their primary language, whereas 54% said that spanish was their primary language.

Moreover, only 76% of the population _understood_ valencian, only 56% was able to speak it, and just 25% claimed to be able to write it correctly.

The numbers today are probably even worse.

BTW, road signs and advertisements are written in both languages almost everywhere (just like in Catalonia, or any spanish region where there is more than one official language).

[1] http://ca.wikipedia.org/wiki/Valenci%C3%A0

I can confirm this. I was born in the Valencian community and I was "forced" to study Valencian language at the school as part of a "normalisation" process so the language wasn't lost.

When I was working as OSS consultant in Valencia one of the main issues was the translation of OpenOffice to Valencian language because for political reasons (mostly) the Catalan version couldn't be used. Let's say it is complicated. To cut the story short: some people in Valencia don't want to accept that Valencian is dialect of Catalan and they insist that it is a completely different language (total BS, but politics isn't about science, is it?).

Yes, some words are different, like en_US is different from en_GB; but the whole absurd thing got to the point that they didn't want to use the locale for Catalan (ie. ca_ES@valencia or similar). Again, because political reasons they wanted to use va_ES but then OpenOffice couldn't integrate the translations in their source tree because va_ES is not an accepted locale in the IANA database.

It was a total mess and it meant that they were always playing the catch up game translating again and again OpenOffice with each new version (I don't know how things are now in LibreOffice, but OpenOffice was a total nightmare).

All together is a little bit sad to see how resources are wasted.

It is a bold exaggeration to say that catalan is as different from spanish as portuguese or french. It is definitely a language, but extremely similar to spanish, much more than any other. Valencian resembles spanish even more.

While you're quick to conclude that valencian is getting extinct, don't forget about all the areas in Valencia where valencian was never, ever spoken, but it was forced into schools, and laws that require an official degree in valencia to work as a government official were set up.

Catalan at source is probably as different to Spanish as French or Italian are. But probably, centuries of stronger influence have induced more similarities: most people in Catalonia are bilingual.

Especially for those who don't speak either:

Spanish: "Hay muchas cosas", "Tenemos una vez" French: "Il y a beaucoup des choses", "Nous avons une fois" Catalan: "N'hi ha moltes coses", "Tenim una vegada" Italian: "Ci sono molte cose", "Abbiamo una volta"

I think it's a quite representative example of the kind of differences & similarities.

"Hi ha moltes coses". That "N'" is a common error. That "Tenemos una vez"/"Tenim una vegada"/etc. doesn't really make a lot of sense to me, but it's technically correct.

Portuguese: "Há muitas coisas", "Temos uma vez"

I'm Portuguese. I'm fluent in French, and minimally capable in Italian. Without formally studying Spanish, I can understand spoken and written Spanish, as well speak it (albeit poorly). I can't understand most words in spoken Catalan, and can't speak it altogether.

Catalan is a close enough language that I can understand its written form, but it is not as close to Spanish as Spanish is to Portuguese.

> It is a bold exaggeration to say that catalan is as different from spanish as portuguese or french.

Sources? Being a lazy guy, I turned to wikipedia [1], where it explicitly says:

> Despite being mostly situated in the Iberian Peninsula, the Catalan vocabulary exhibits marked lexical differences with Ibero-Romance (Spanish and Portuguese), instead showing general affinity towards the Gallo-Romance group.

The Gallo-Romance group includes French and Italian.

About regions where valencian was never spoken and forced into schools, what is your point exactly? I just said that valencian usage is becoming less frequent lately, and the numbers support this affirmation. Are you trying to say that, with the current laws, the language can not become extinct? Then I may agree with you, but laws can be (and have been) changed...

[1] http://en.wikipedia.org/wiki/Catalan_language (lexical comparison section)

Edit: missing link.

Just because it has (obvious french) Gallo-Romance influences, doesn't mean it's less similar to spanish than poruguese.

My point is that you're complaining that it's being used less, while it's actually being forced onto places where it simply has never been used.

I reply here because I can't reply to the below comment.

The claim in wikipedia says that it has more general affinity to the gallo-romance group than to the ibero-romance group. I may be wrong, but I read that as "it is more similar to french/italian than to spanish/portuguese".

Also, I'm not complaining about anything (and I couldn't care less about wheter valencian becomes extinct or not). I was just refuting the op affirmation statement that "You will mainly hear Catalan in the streets of Valencia, not Castillian Spanish".

>* The claim in wikipedia says that it has more general affinity to the gallo-romance group than to the ibero-romance group. I may be wrong, but I read that as "it is more similar to french/italian than to spanish/portuguese".*

There might be a terminology mismatch here and you are talking about different things.

Perhaps the core language (syntactic structure, regular verb conjugation patterns, core words) is very like Spanish and that comment on Wikipedia is referring many non-core words in the common vocabulary being pulled from languages common in other neighbouring areas.

Caveat: I am no linguist nor do I speak anything much other than Correct English and American English.

"It is a bold exaggeration to say that catalan is as different from spanish as portuguese or french."

What?

indeed!

Valencian (Valencià) is not a version of Spanish, but a different language. It's the same language that is spoken in Catalonia (and which there is called Catalan), and in the Balearic Islands (where it's called Balear).

It's amazing the amount of languages that are everywhere...

> In my experience, the main cost to adopting F/OSS for office software is interoperability with people sending word and excel files, not support or training.

Couldn't this be handled with WINE or a virtualizer to run Word and Excel?

> Note, LibreOffice is available in the native language of the users, Valencian, and has spell-check dictionaries for that language. Presumably, the Valencian government puts out most communication in Valencian, so there may actually be savings in not having to develop and maintain their own Valencian dictionary for Office, which Microsoft does not provide, and reduced support as the interface uses terms more familiar to the native Valencian speakers.

If they're already in the position of having specialists do translation, or having an entire office that's somewhat bilingual on the computer, running Word and Excel virtualized might not be such a burden comparatively speaking.

What you call "costs" actually are salaries/jobs for people living there -- unlike fees paid to MS.

This is an insightful point that those who push for austerity measures and privatisation always seem to miss.

A government saves money if it pays €x to employ local taxpayers instead of sending that same amount abroad. It will still break even on the decision if it employs locals for €x+n, where n is the amount they recoup in tax from those locals. Even more if those locals would otherwise be unemployed.

Government gets its money by siphoning it out of the local economy, making that capital unavailable for private use while creating nothing of value. Government jobs are a net drain on an economy.

This. Even if a great deal of the savings from cutting MS fees is forwarded to training the current/new tech support it's still a win.

Always the same training story. What is the percentage of microsoft office users who has received a training ? I do not know any. How many trainings have been given when migrating to the latest releases of MS O that are so different than the usual interface ? The deployment of OO instead of MS O is not more complicated than an upgrade of MS O toward another version of MS O. And for the support costs, the main problems of MS O were concerning compatibilities between releases. These issues do not exist for OO. The only way OO can be more expensive than MS O is if many people ask to have MS O in addition to OO.

Training is ongoing and already handled by their budget and support costs are probably the same.

This is pretty much just comparing the license cost.

Edit: we have a client who has 2000 users and they have two people dedicated to managing just licensing and have to operate KMS servers etc. That should be factored into the cost as well.

Yep. 1.5 million euro / 120,000 computers = 12.5 euro per machine per year.

That's probably negligible considering costs of installation, productivity, support, etc.

(I'm still happy to hear about this, but I admit that it's more ideological for me than economic.)

But re-imaging desktops and re-training can largely be internal costs. It's asking your staff to add another item to their to-do list. Whereas, keeping Microsoft licenses means actual cash being spent.

Not sure what you're suggesting. A cost is a... cost. Whether it's a loss of productivity or a license cost, attributable to OPEX or CAPEX, it's still a quantifiable expenditure. If a (any) cost isn't quantifiable then I'd suggest there are bigger problems than choosing between proprietary and non-proprietary.

If the cost is attributable to the use of capacity that would otherwise be left unused (idle workers), then the marginal cost is effectively zero, even if it shows up as nonzero in the books.

dont get carried away mate - these are government employees. They typically dont work beyond 3 pm anyways - 30 minutes more to learn how to use Calc isnt killing anyone. Besides - its about long terms savings. They will save this much every year. Over time it adds up. Especially considering spain is in financial trouble right now!

> Whether it's a loss of productivity or a license cost, attributable to OPEX or CAPEX, it's still a quantifiable expenditure.

Yes, but some kinds of costs are still easier to hide.

The salaries for those people are real costs. If you don't add things to their to do lists, you can hire fewer people and pay less.

It's all cost.

There's usually some float in people's plans - if there isn't you tend to go the other way and have projects overrun.

Where does the money go to retrain the 100,000 users? And who gets the money to re-image the machines? It goes into the local economy.

Now, where does the money end up when you buy 100,000 copies of Office? The United States. Perhaps some offshore bank account where Microsoft parks their cash next to Apple and Google's billions?

> - Re-training the (I assume) 100,000+ users?

Whatever it was, it was well worth it since in the future everyone will use open source software.

What's the alternative? Keep paying more for Office software every year, while risking having NSA gathering all those documents?

Our business side uses MS Office simply because the rest of the world uses Office and we frequently get sent Excel documents. Engineers use LibreOffice and boy oh boy, it's a tremendous pain when the business side sends over Excel documents of reports that they would like to get done. Even Word documents are formatted correctly and it's been a painful experience overall. Sure, they save money from the software licences but what about the poor experience and time wasted? I agree that I don't believe that F/OSS is cheaper. There's a lot of costs that don't get quantified in the analysis.

You are missing a point though: this is the public administration side. They can (and should) demand any document sent to them to be formatted in a public open standard (such as ODF).

If anyone, it's the companies interfacing them that will have to bear those "interoperability" costs. At least until they also switch to LibreOffice or something compatible enough. IMHO this is much better than them requiring Office-formatted documents.

I agree that we should demand an open public standard but essentially it is still a cost nonetheless. We're just pushing around the costs to other companies instead of absorbing it within our own company. I think the government does stand a chance in demanding which standards they will accept though but for small companies looking to do business, it's gonna be a challenge trying to get customers, suppliers, partners etc to give you a format that you specify.

I totally agree with you, but this is a chicken-and-egg problem. Unless someone starts demanding such standards, the challenge will always be there.

Letting the administration be the pioneer in this front is sensible both morally and practically. Morally because the administration should be pushing for the greater good for all citizens, and practically because people/companies can not just "choose another provider and sidestep the issue".

Once nearly everyone is forced to have processes in place to interface with the administration using open standard formats, small companies looking to do business will have a much easier time asking for these formats too.

I moved from Office and started using LibreOffice several years ago and still run into issues where people can't open the LO docs I send over - even after I've saved in the MS Docx format.

I can only imagine the headaches they are about to incur with this switch.

Have you reported the bug ? Can you provide such an example, or is it FUD ?

Sometimes "I can't open" means, I opened it, and the formatting was changed, so now I'm telling you this...

In this case, the flaw is not with the software, but a bad document format with a closed standard, as well as an undefined and misdefined standard of how to render the documents.

True. I think you hit the nail on the head of the "undefined and misdefined standard of how to render the documents". If the documents are completely interchangeable between MS Office and LibreOffice, I would buy the cost savings a little better but because there's a loss of productivity from dealing with incompatible formats, I would argue that it's not just a straight forward swap of software and the cost of the software alone.

Yes, agreed - except for the longer term, the practical repercussions are the same, whether the software or the standard is to blame (philosophically, almost). But it's a distinction that is important to keep - for me - not least to learn from.

[0] claims that the city of Munich saved about 10 million Euros by switching to their own Linux distribution ‘LiMux ’ [1]. So, yes, F/OSS is cheaper.

[0] http://www.h-online.com/open/news/item/Linux-brings-over-EUR...

[1] https://en.wikipedia.org/wiki/LiMux

this seems to be a trend: http://www.ubuntu.com/products/casestudies/french-national-p...

Good stuff. ^_^

What if you count the exit costs of vendor lock-in as costs associated with that vendor? Suddenly Office is a lot more expensive.

"TCO should include exit costs and cost of vendor lock-in" https://joinup.ec.europa.eu/news/tco-should-include-exit-cos...

M$ fanboys spreading FUD?

Let assume that the 100.000 users all use Office for 2 hours a day on average.

Now let's assume that MS Word will give you a 1% performance increase over LibreOffice... just one percent, that's 72 seconds a day. 72 seconds spent wondering "how do I make a pivot table in Calc like I do in Excel"? or "How do I make a table of contents in LibreOffice Word"?

Now let's assume the average pay for these people is 15 euros/hour. those 72 seconds therefore cost 0.3 euros per employee.

Now multiply that by 100.000 users * 270 working days a year and you have 8.1 million.

Congratulations, you just lost 6.6 million euros.

... Sorry FOSS/Linux evangelists, but apparently none of you can do math.

You are pulling numbers out of your ass and claim we can't do math?

Give us some citations to document performance loss, and further citations do document that the performance loss is sufficient enough to reduce the amount of work the staff actually carries out (if they're less productive per unit time spent working, but still completes the same amount of work, then the cost is not borne by the government - it is fairly normal for workload to be quite elastic), and we can talk about your numbers.

You also don't account for another factor: Even if there is a loss, it is not given that it is a net negative: License payments to Microsoft leave Valencia, and Spain. Payments to any additional training and support staff or additional assistance for LibreOffice is likely to stay in the local economy, and part of it will find its way straight back in in the form of taxes.

It is perfectly possible that they've messed up the numbers, but you've not presented anything resembling a case for that.

You might also note from the article that this is not Valencias first large scale migration - they claim to have saved 30 million so far from switching the regions schools to Linux, so chances are they actually have better numbers than the ones you've made up.

I made these stats up in about 30 seconds. I'm guessing they'r e not that far from the norm. Even if the average use was only 1 hour, or 30 minutes, and the average pay for half of what I assumed, they should still be losing money.

I just wanted to point out, people forget these things when they do a cost/benefit estimate.

But you've made an assumption that they will be more productive in MS Office, with nothing to back that up. I've see users (particularly non-technical, "basic" users) be more productive in LibreOffice. So they may have messed the numbers up and be saving even more than they thought.

You're full of assumptions and guesses, but little facts, and you're assuming Office is more productive when it could easily be the opposite.

Because nobody needs retraining or relearning when moving from one version of MS Office to the next.

>let's assume that MS Word will give you a 1% performance increase over LibreOffice

That's a pretty big assumption.

I've used both products. I didn't notice any noteworthy differences when it came to regular usage. I actually think the new-ish Ribbon UI is extremely confusing and slow to navigate.

This certainly isn't like Gimp vs Photoshop.

Ok. Let's try some more imaginary maths. If Valencia has five million inhabitants, and one in ten in these, in turn, send one document to the regional government every year, that would be, at a hundred Euro each, about 50M€ going to Microsoft. If the government uses LibreOffice, that's a big saving for the people.

On top of that, I guess most of the 100.000 users at the government -doctors, teachers, judges, clerks...- have never used a pivot table, and won't get those extra 72 seconds per day. This ain't Henry Ford's factory and there are much bigger externalities to take care of before this kind of imaginary optimizations.

Imagine how much those tea breaks must cost them then and going for a piss.

BILLIONS I tell you BILLIONS!!!

Give them IV drips and buy them diapers quick!

> let's assume that MS Word will give you a 1% performance increase over LibreOffice

If we would assume that the new ribbon-UI of Word is unfamiliar to people, then it actually could be that LibreOffice is easier, more performant, for people accustomed to (old) Word, than Word itself.

The ribbon is far more productive as it is at least discoverable without much effort.

Most people have no idea what the toolbar buttons do. To find out you have to play mystery meat with tooltips for ages.

We have a mix of Windows machines with MS Office and Ubuntu machines with LO here for our admin staff (with each staff member using both for different tasks), and I get far more questions about how to find things in the MS ribbon than I do in the LO menus (and LO tends to be used for more complicated operations, it part of a high-speed scanner / data work-flow system).

> Most people have no idea what the toolbar buttons do.

But one doesn't need to use the toolbar buttons, one can still select things from the menus in the old fashioned way.

Can still do that with the ribbon. Press Alt.

Most people have no idea that that is possible. To find out you have to play mystery meat with keys for ages.

In my student days, "mystery meat" was for stuff that you never found out about what it was.

RTFM, it's even searchable these days.

Or, you know, learn/set up keyboard shortcuts if you're going to be working with the software regularly.

Why are you multiplying it by the days in a year? Are you seriously saying that every single user will waste 72 seconds every single day when they use LibreOffice?

1. Regarding productivity tools, people tend to do roughly the same thing at work each day. What they type in might be different, but what they want the software to do doesn't differ that much.

2. People tend to be able to remember things from one day to the next. If I start using a new IDE, I don't have to spend an extra minute every single day for a year looking for the "build" button or shortcut. A user who works out how to format a line as a heading in LibreOffice won't sit around the next day, confused about how to format a line as a heading.

3. Regarding most commonly used functionality, the UI of LibreOffice is pretty intuitive for someone who has used any kind of office productivity software.

4. Regarding rarer functionality (e.g. Pivot tables) I would be surprised if anywhere near 100000 of those 120000 users ever use them. Most people just use a spreadsheet as a place to put tabular data, many don't even do simple calculations like summing with it. Most people just type big globs of text into word processors and do some simple formatting. Those who do use less common functionality tend to have the ability to work it out pretty quickly and either remember it or write it down.

As many have already said. The unfamiliarity problem you describe is the same with a new version of MS Office. When I got a new version of MS Office, I spent some time trying to find features that I used a lot before, but once I found them, I was back up to speed. If the timing of the change corresponds to the need to upgrade MS Office (e.g. existing version going out of support), then the disruption would be pretty much the same.

"How do I make a pivot table in Excel X like I do in Excel X-1?"

People will need help adapting to the new software whether it's FOSS or expensively licensed. Unless your suggestion is that they stay on Office 97 indefinitely, those types of training costs will always come up.

It's like saying you wouldn't use cars that run on water because you'd have to change your car fleet.

It's an investment. The question is not how much you save this year, it's when will you actually save money.

Reply on news.ycombinator.com