Hi all,
This is a short summary of some of the problems that I see
with LibreOffice, and this is written with my personal / Collabora hat
on. People are welcome to question my motivations - but my mission is to
try to nurture a successful FLOSS project that creates excellent FLOSS
Office Productivity software and makes it freely available to all. Many
here will share that goal I hope.
Nevertheless there are some big problems currently. Perhaps
you think you have a neat solution to one of them. I'd love to hear
about it - but solving or obsessing about just one is unlikely to do
the job:
* LibreOffice is at serious risk
Frustration with how TDF markets and positions its 'product'
(LibreOffice) against the ecosystem that contributes the majority of
the coding work is at an all-time high. That ecosystem itself is under
long term stress.
Despite years of patient work, writing up the problems here,
talks at conferences, personal pleas for change and improvement, and a
number of tweaks, nothing -effective- has happened. You can read about
the situation here:
https://people.gnome.org/~michael/data/vendor-neutral-marketing.html
* That's too long (despite the pictures); what are the problems ?
Read the Ecosystem / Sustainability minutes from our board call.
https://wiki.documentfoundation.org/TDF/BoD_Meetings#Minutes_2020-05-22
I've helpfully appended it to this mail. It has the history of
the ecosystem to today presented by myself & Thorsten as bullets. Then
we have some of the ecosystem problems:
+ how to differentiate from LibreOffice:
+ support - but why buy that ? it's great.
+ how to differentiate inside the ecosystem ?
+ proprietary bits suck badly, obviously.
+ how to get the message out that it even exists ?
+ for less than the cost of the software.
+ how to get the message out that it is more authentic and
genuine to get LibreOffice from those doing the majority of
the code contribution, than the free version from TDF ?
+ how to build a brand that stands for quality and
support vs. the Goliath brand: LibreOffice ?
+ how to make LibreOffice not mean "everything for
free, please don't pay anything" to most users ?
Anyhow - more details and some FAQ below:
* Surely companies have to buy support & security updates ?
They always complain to me about the lack of support wrt.
avoiding using FLOSS !
Sadly no. Microsoft gives poor to non-existent support to the
majority of users so ~no-one expects to buy it, they expect to buy a
product. Enterprises tend to test a version & deploy it to their
desktops and leave it there - they can do that with LibreOffice from
TDF.
It is routinely the case that I meet organizations that have
deployed free LibreOffice without long term support, with no security
updates etc. Try the Cabinet Office in the UK (at the center of UK
Government), or a large European Gov't Department I recently visited -
15,000 seats - with some great FLOSS enthusiasm, but simply no
conceptual frame that deploying un-supported FLOSS in the enterprise
hurts the software that they then rely on. Or a giant Pharma company
in the news right now; companies do it left & right.
This became a familiar problem when after the OpenSSL /
heartbleed debacle it was discovered that just a couple of people were
part-time maintaining something vital to the whole world's internet.
This is an extraordinarily common pattern, people come to tell
me how many free seats they've installed in large enterprises - and
while this is a triumph; they tragedy is that they stop at this point.
Far too often the whole thrust of the selling was "zero cost" - which
is a terrible way to market FLOSS. They are now used to downloading
Chrome or Firefox and deploying these advertising supported products
for free everywhere. Building our USP as zero-cost is a horrible way
to market LibreOffice to enterprises.
So - lets turn this around - can anyone thing of more than
five enterprises that paid for support or instead (just as good)
contributed meaningfully to LibreOffice instead ? Munich, and ...
Of course we maintain and promote lists of enterprises that
deployed for free with no support ?
=> It is the norm to deploy LibreOffice from TDF in
enterprises, and pay nothing for support &
maintenance that can go into development.
+ its that good.
* You're too expensive: I can get cheaper support from LXYZ instead ?
Another pathology is that there are companies who ship
LibreOffice, often claiming support, but then file all their tickets
up-stream and hope they are fixed for free. Naturally they are cheaper
in government tenders, they use our brand, they leave the customer
with hundreds of un-fixed bugs, and all of the users with a terrible
experience.
These companies also seem more 'genuine' since they call their
product 'LibreOffice', as compared to those who actually contribute
who try to build their own brands.
This is not a small problem; there are many millions of users
in this situation, there are multiple companies I know of with scale
of $10bns in annual revenue, to Linux distros, to local service
companies who do this.
=> The LibreOffice brand is devalued and we have no way
of telling people that the product they deployed was
not suitable for deployment in an enterprise and has
no effective support.
* but we tell people to get support & services already!
Sure - there is some text on the download page (one of the few
pages with hits) - but it doesn't work, you can track the rates of
conversion, they are extraordinarily low before thet get off TDF. We
also mention this in our release notes helpfully, again - not
effective.
* Surely if we seed the market by giving apparent enterprise
products for free - people will pay ! just give more away !
Ten years in, we notionally have 200m deskop users - but we
have extraordinarily few paying for support & services. The conversion
rate is at the 0.01% level - not a million miles from the 0.05% level
that we see for people visiting the "professional support" page on the
TDF website.
Even if we converted the whole world to LibreOffice on the
desktop, if the everything-free-for-everyone messaging continued - it
seems unlikely that we would be able to improve and sustain the
software.
=> emphasizing gratis from TDF over libre as we do
~everywhere probably screws up our ecosystem.
* Is it really so bad ? everything seems fine ...
Words fail me to express how beyond-utterly-broken the
existing TDF / desktop model is for the ecosystem around selling
Desktop LibreOffice.
Collabora - despite C'bra still putting a lot of work into
LibreOffice Desktop, having an outstanding support capability, doing
lots of marketing, being the largest code contributor to LibreOffice,
and having lots of existing happy customers / references for desktop
LibreOffice, ... etc. etc.
We have not had -one- -single- -new- Collabora *Office*
customer since 2018 - zero.
=> so it makes no economic sense at all to invest in
-Desktop- Libreoffice you will never see a return.
That is manageable - we are investing heavily in creating
Online and that is going well, and it funds our work on LibreOffice.
As an aside - what precipitated this new attempt to try to
sort this all out for me - is some on the Board wanting to bring this
known to be destructive for the ecosystem model of branding including
suggesting it is suitable for everyone without anyone paying,
competing with the ecosystem that creates 95% of Online, as well as
crushing our brand and hence lead-flow with the LibreOffice brand.
Checkout the appended ecosystem history, this is a conclusion
our ecosystem has discovered left & right.
Why does LibreOffice Desktop not have XYZ cool new feature I
badly want, contributed by companies: this is why.
* Other ideas: what about crowd-funding ?
Various kind people have pointed users systematically at
crowd-funding sites in bugzilla and elsewhere. This has worked quite
well in the past for one or two features - but for more general or
smaller improvements has failed. I believe after some years we had an
~Eur 200 bounty for implementing an Online collaborative version of
LibreOffice put up vs. the Eur millions of aggregate investment this
has cost.
So - focused crowd-funding for features can work - if there is
TDF marketing support particularly.
* But surely if everyone donates to TDF and all is well ? we
just need to ask for more donations !
Currently TDF sits on a ~Eur 1.5 million cash pile yielding a
zero to negative interest rate.
For various complex, structural, organizational, legal
etc. reasons TDF has managed to effectively tender only one smaller
partial feature (related to ODF 1.3) since the ESC and Board voted for
community submitted items from 2018.
However - even if TDF spent its existing budget on feature
development it would contribute around 10% of what the ecosystem
currently puts in.
Possibly if we give TDF 10x more more money - it will become a
more dynamic organization (though still run by a committee of ten);
perhaps that is possible.
* These naming changes seem to suggest some people need to pay ?
I won't be able to install it in XYZ enterprise if called Personal
This is in large part a feature. Clearly people can use it for
free, just as before - but they would occasionally see the splash and
thing: "I wonder why it says Personal" (or another tag) there.
Probably this nudge alone is enough to try to encourage real
contribution to LibreOffice, and get the numbers of users buying
support and thus contributing, or else contribuing themselves up from
~zero.
* Can't you differentiate by having proprietary add-ons ?
I'm an old free-software hacker. To me proprietary things are
at best really unhelpful. The great goodness of having mild
differentiation and steering of people to buy support in the product
name, is that it avoids proprietary differentiation. You may recall
Sun/Oracle did this in the past with StarDivision and it was not much
loved. Also - the logic of this 'open core' approach is clear: invest
mostly in the non-open, non-core bits, which is bad for LibreOffice.
* Why change anything now ?
As we look around the industry we see tons of organizations
exploring ways to solve similar problems. LibreOffice's is
-particularly- challenging, because we aspire to being a vendor
neutral project. There are reasonably well-known ways to build a
company controlled, branded, FLOSS project - we know and love lots of
them: openSUSE, Fedora, Nextcloud, ownCloud etc. this is the
norm. With TDF we tried to do something far harder - to create an
vendor neutral ecosystem that can help retain the community spirit
while delivering on our mission. That has proved extraordinarily
harder.
The particular timing was precipitated by the demand to bring
the long-term proven zero-sales, zero-investment model to Online, and
to do so very rapidly for LibreOffice 7.0.
This proposal was a constructive one to try to take the
potentially catastrophic risks associated with a single model and try
to ameliorate them by in a single step encouraging people to pay for
something, that can be re-invested.
Perhaps having an 'Edition' in itself, while a tiny nudge is
enough regardless of its name to encourage people to make people ask:
is this right for me in an enterprise.
We can go slow, or do something else - it is entirely possible
that something else works. For my part I would really appreciate that
we as a community test whatever marketing improvements are planned on
the desktop product (which can hardly get worse for the ecosystem), to
see if they are even slightly effective. Then as/when/if this works we
can apply it to other sub-projects like Online, when we know the
impact on the ecosystem.
If we want a pure volunteer project, with no significant
ecosystem behind it we can see what that looks like over at the Apache
OpenOffice project. There is integrity in that of course, but is that
somewhere we want to risk going ?
For my part - I'd like to try to work with others to
understand all of the motivations, and to somehow, together chart a
path towards a better way of marketing and positioning LibreOffice and
its ecosystem - as Italo has outlined. Many of the above issues are
significantly addressed in this proposal - and I think it forms a
great basis for discussion, hopefully its possible to map many of
these solutions back to the problems I outline now.
---
* Excerpts from:
https://wiki.documentfoundation.org/TDF/BoD_Meetings#Minutes_2020-05-22
2. Inform/Discuss: Ecosystem & Sustainability (Michael & Thorsten, All 15min)
+ Board wants to be more transparent about decisions
+ that is good, but we need to publish more data.
+ please listen, absorb - everything we say will be
in the minutes or published later:
+ questions / challenges possible at the end.
"The foundation promotes a *sustainable*, independent and
meritocratic community for the international *development* of
free and open source software based on open standards."
LibreOffice has included companies into the ecosystem from the beginning
+ the word 'community' includes people working at companies
+ inspire a higher loyalty to the project & the idea: unifying.
+ economics seems harsh
+ value volunteers time & support of course.
+ need to model & understand these things - Eco-system
+ When we started:
+ Sun -> ~50 developers -> we were eager to include them
+ IBM -> ~10 developers -> tried to get them to work together with us
+ big happy family...
+ Oracle -> OpenOffice not a contributor to net margin -> gone.
+ IBM -> encouraged us to form, didn't join -> differentiation.
+ we all lost hard.
+ Then: Linux Distros
+ SUSE - 15 developers, RedHat -> 5 developers, Canonical - 1
developer
+ Munich? -> peaked at 7 - now down to ~1
+ The flourishing years 2012-~2014
+ KACST ~5, CloudOn ~2, Igalia ~1.5, Lanedo ~2,
Ericsson ~2, Synerzip -> ~12,
MultiCoreWare -> ~15 (2 months)
+ all of these disappeared: no economic driver.
+ More recently (Thorsten)
+ SUSE - 15 -> 0 devs. -> 2013
+ Canonical from 1 to 0 devs. -> 2017
+ RedHat from 5 to ~2 developers - 2018
+ can't assume Linux Distros would magically do the coding
+ Now we are here:
+ 1&1 -> 1 developer
+ Munich -> 1 developer
+ TDF -> went from 1 to 2
+ RedHat -> ~2 developers
+ RedHat -> 2 developers
+ NISZ -> 3 developers
+ CIB -> 7 developers
+ Collabora -> 25 developers
+ and some more volunteers.
The ecosystem is not growing, but not shrinking ~40 paid developers
+ very little virtuous cycle driving growth in sales -> feature /
function
What does the ecosystem do ? (Michael)
+ ~70% of commits from ecosystem companies
+ tend: larger features, longer term development, maintenance,
GSOC ++ mentoring
+ ~30% of commits community
-> but ~70%+ of mentoring is companies.
+ look at how a project looks like with volunteers but no commercial
ecosytem:
+ Apache OpenOffice -> good people ...
What does it look like in money ? (Thorsten)
+ easy to work out from our annual report
+ estimations ->
+ avg. TDF staff member FTE cost order of Eur 50k without
overheads
+ So 40 people at TDF would cost ~Eur 2 million
+ TDF spends ~Eur 150k on feature development / tenders
outside of fixed-costs per annum currently
+ do it in-house:
+ we could pay 3 people with that
The ecosystem provides a 10x multiplying factor for development
=> TDF is the tip of the iceberg that is the ecosystem.
=> We cannot meet our purpose or sustain the product without a healthy
ecosystem.
Compromises to drive the ecosystem / pull money into it & TDF: (Michael)
- no LTS - short, frequent release cycles
+ neither of these seem to deter mass use in business.
- nag dialogs, info bar stuff -> to drive TDF donations
+ community members unhappy ... -> professional product.
- tell people software is un-supported when inappropriately
using Online - important for TDF to have users visiting
download page, also for downloading security updates for
LibreOffice
+ live updates provided by community as a patch / feature.
+ not enabled.
+ we know: when we don't update the "new version" XML
+ donations drop substantially
+ TDF is funded by people visiting the download page
=> Example -> providing free LibreOffice in an app-store
+ self-updating, no need to visit TDF ...
+ sounds wonderful - but potentially deeply problematic for TDF
& sustainability.
+ needs to be looked at -very- carefully.
Thorsten - People sometimes confuse paying for a product, when
donating (c.f. feedback we see on Ask and Bugzilla)
+ normal feedback "I paid, but can't download"
- TDF download donations sometimes believed to drive feature development
+ implement feature with money.
Desktop is quite different:
- Online - currently scarcity: source code, only development binary
dockers of online
- CIB & C'bra charging for LibreOffice in app-stores & re-investing -
perhaps others.
The "non-contributor" problem
+ Developers:
+ Munich -> initially didn't contribute to the code -> then saw the light &
joined us
+ Linagora -> no significant contributions from whole French Gov't.
+ Brazil -> 1 million users -> no paid developers until recently.
+ others: RedHat/Collabora/CIB - contribute all their code back.
+ work as members of the community.
Conclusions (Michael)
+ why is ecosystem success seen as failure ?
+ why is the ecosystem 'othered' as non-community ?
"The foundation promotes a *sustainable*, independent and
meritocratic community for the international *development* of
free and open source software based on open standards."
"The only constant is change" (Thorsten)
Change is inevitable, and necessary -but- those who advocate
for change without careful consideration of the consequences
must bear responsibility.
Need to assess, and balance - with the other compromises
listed if parameters change substantially, we need a plan to
substitute for what we're certainly loosing then in terms of
ecosystem contribution.
Lets not loose again after loosing Sun, IBM, Linux Distros ...
--
michael.me...@collabora.com <><, GM Collabora Productivity
Hangout: mejme...@gmail.com, Skype: mmeeks
(M) +44 7795 666 147 - timezone usually UK / Europe
--
To unsubscribe e-mail to: board-discuss+unsubscr...@documentfoundation.org
Problems? https://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: https://wiki.documentfoundation.org/Netiquette
List archive: https://listarchives.documentfoundation.org/www/board-discuss/
Privacy Policy: https://www.documentfoundation.org/privacy
