LibreOffice 6.2 is here: Running up a Tab at the NotebookBar? You can turn it all off if you want While Microsoft was upending the scorn bucket on its own productivity tools, The Document Foundation shoved out an update of LibreOffice, replete with user interface tweaks and improved Office compatibility. Originally forked from OpenOffice back in 2010, LibreOffice comprises a word processor in the form of "Writer", a …
> it wanted to shut down several running programs
Windows file systems have a limitation that a file cannot be updated while a process has it open - not without causing a crash. This is due to the directory entry having both the filename and the pointer to the data area on disk. On *nix the i-node system caters for updating an open file because the directory entry has the filename and an inode number. The file than is open can continue to be used through its i-node while the replacement file can use a new i-node and new data blocks. The old file will be deleted when all processes have closed it.
This is why Windows requires a reboot. Updated files are given a different name and a script is created to delete the old files and rename the new ones. This script is scheduled to be run during the next reboot.
>........This is why Windows requires a reboot. Updated files are given a different name and a script is created to delete the old files and rename the new ones. This script is scheduled to be run during the next reboot.
And this is why windows is a bag o' shite
And this is why windows is a bag o' shite
If it were JUST THAT, then Windows would be almost useful, Windows has more problems than I could concisely enumerate, given the remainder of my lifetime.
The main issue which is the mother of most of its issues, is that it is proprietary software.
> Still clinging on to your 32-bit Linux installation? The team warns that after this version there will be no more Linux x86 builds.
No surprise there, major Linux distributions have dropped (or soon will drop) 32-bit builds as well. I expect soon popular software will be so full of "all the world's 64-bit" assumptions that keeping compatibility with 32-bit gets too laborious. Bit like porting software from 32-bit Unix computers to MS-DOS was a major pain back in the past.
Except unlike Windows, there is no penalty in speed or code size running 64 bit Linux on Atoms (with 2G RAM) that MS only distributes 32 bit Windows 10 on. Also Linux 64 runs 32 bit programs better than 64 Windows and sometimes Windows 32 bit (on WINE) that won't run on even Win7 64bit.
Also 64 bit Linux fine on 32G Flash / SSD /HDD, which is getting to be a problem for even 32bit Win 10.
Likely there will still be 32 bit Linux for routers, set-boxes and Gadgets etc where 64 bit is too much expense and power.
I found that Debian can install on those in 64 bit, but for Mint you've to copy a file to the USB stick because the Mint folk (quite reasonably) think why would a 64bit CPU have a 32 bit UEFI.
Also a place in hell for whoever in Intel decided to have a 64bit CPU with PHYSICALLY only 2G RAM addressing. I don't believe the package couldn't have had 1 (4G) or 2 more (16G) pins. Crippled from birth.
Or tablets & laptops with soldered in ONLY 32G Flash/SSD after Windows 10 launch?
Well, at least such HW can support 64 linux well enough, though if it's rotatable and has touch you may need a script to handle touch in portrait.
Running Linux Mint Mate desktop on a 2GB Atom netbook, with Chromium, Thunderbird and a couple of terminal windows open, takes 774MB of memory with the 32 bit version, and 1.2GB with the 64 bit version. The 64 bit is faster, until you've been using the browser a while and it starts to swap, the 32 bit version lasts a lot longer before it starts to do this. The machine has a SSD, but being an Atom is a slow SATA2 bus, so swapping is still very painful.
I've got both versions, as already some 3rd party software is x86_64 only. So it looks like ARM is the only place where 32bit is OK, and incidentally a 32bit Raspberry Pi 3 runs faster and lasts longer before swapping, despite only having 1GB of memory and driving a 1920x1200 monitor rather than a tiny 1024x600 laptop screen.
"These aren't even obscure distros either."
Debian Stretch for one. That's because they're Just Works rather than Only Just Works distros. They run versions which were sufficiently mature when the release was frozen. You can uninstall the standard version and install a more recent version if you like. Alternatively you can keep the standard version just in case and install a more recent one alongside. In Debian that goes into /opt. I've just removed 5.2 from this laptop having installed 6.1 alongside 6.0. Having 3 versions seemed like overkill.
Well, actually native 64 bit code ensures 32 bit cleanliness. You don't have the opportunity to use upper "unused" bits for flags'n'stuff and other nasty hacks.
The major reason is that "stuff doesn't fit in 32 bit"
For example, Kerbal Space Program dropped 32 bit because textures and object models were getting too big to fit, especially with all the hi-res art rework going on.
(The funny thing is that KSP was the sole reason that I converted my Debian box to 64 bit in 2013, because it ran so much better than all the other versions, including the Windows versions.)
Well, actually native 64 bit code ensures 32 bit cleanliness.
Not quite, I'm afraid. On 64-bit C compilers in Linux, "long" is a 64-bit type. On 32-bit compilers it is a 32-bit type. A developer working only on 64-bit systems will soon assume the 64-bit "long", instead of laboriously writing "long long" which would have the same 64-bit size on both systems.
(I believe 64-bit Windows handles this differently, and keeps "long" 32-bit for just-in-case backward-compatibility, which I think is wimpy).
"On 64-bit C compilers in Linux, "long" is a 64-bit type. On 32-bit compilers it is a 32-bit type... 64-bit Windows handles this differently, and keeps "long" 32-bit"
I would (altogether subjectively) expect that on a 32-bit compiler, int would be 32 bits, also int and long could never be the same size. So that's weird. Was int still 16 bits (back then) or something? And what's a 32-bit compiler anyway, an old version of a compiler (ca 2000) that cannot yet generate x86_64 code? Which compiler? Did they (Intel, GNU, etc) agree on when int would graduate? I know, I know. STFW. My point is, that kinda just raises questions.
By 32-bit compiler, I mean a compiler natively running on a 32-processor, or cross-compiling code for 32-bit processor. On such compilers, "int" and "long" are the same size, at least in all compilers I have used or heard of. (Nothing in the language spec prevents this), On 16-bit systems like the PDP-11 on which C was first implemented, "int" is 16-bit and "long" is 32 bit. Logically, one would have expected "long" to become 64-bit when going to 32-bit processors, but for some reason this was not done. (Probably some backward-compatibility concerns even back then).
The integer size assignment normally used on Linux and Unix systems on 64-bit systems is called LP64, and was chosen by some unix guru cabal back when 64-bit CPUs stated appearing, see http://www.unix.org/version2/whatsnew/lp64_wp.html
I used 32bit Linux for years, then last year I was forced to switch on 64bit Linux is running on faster system with 8GByte memory. I can admit, it does not run the same desktop! workload I had on 32bit, period. The Scheduler and memory management seems to be a fundamental issues, same to them.
" I expect soon popular software will be so full of "all the world's 64-bit" assumptions that keeping compatibility with 32-bit gets too laborious."
Surprisingly, no. It's JUST 32-bit x86 that seems to be being dropped en-masse. If you follow good programming practices and use your size_t, htonl() etc. then you really don't have to worry about word size or byte order. It's JUST x86 32-bit that is being dropped en-masse; those distros that support ARM seem to plan to ship both 32-bit and 64-bit ARM ports for the forseeable future.
When Canonical announced their drop of 32-bit x86 Ubuntu ISOs, they provided quite a good reason; their stats indicated under 1% of Ubuntu users were on 32-bit x86 systems, and furthermore the vast majority of THAT 1% were running 32-bit Ubuntu on 64-bit x86 chips (i.e. they probably downloaded the 32-bit ISO by mistake or because they weren't sure which one to get.) They also pointed out the 32-bit x86 systems are either 10+ years old, or slightly newer low-spec Atom netbooks; therefore they recommend running a lightweight distro (which are generally now doubling down on 32-bit x86 support now that some mainstream distros are dropping it.)
I thank all the developers too. My employer used MS Word, Excel etc for all documents. When a Word document (in .doc format - not the modern variants) or template became corrupted I could nearly always open it with LO and then save it as a .doc . Why LO should be better at opening corrupted MS files than Word itself is a mystery. I am aware that LO does not have all of the legacy cruft that Word has. Why should Word documents become corrupted/unopenable in the first place?
I have fixed corrupted .doc files with Open Office and Libre Office.
I have fixed MS OS problems with live Linux CDs. (Forgotten passwords, Registry problems, virus problems, etc)
I have never fixed any Linux computer with any MS products. I wonder why this is?
"Why should Word documents become corrupted/unopenable in the first place?"
Well, .DOC was at times not a file format in any real sense of the word. If you had "fast save" turned on, for instance, it pretty much would stick a small header at the start of the file, then just do a raw RAM dump of the RAM that would contain your document. Nice and fast, at load time it can just load that RAM dump back into RAM and there's your document. BUT, of course, that means once some newer version of Office had enough internal changes, some lucky bastard at Microsoft would have to work on code to either shuffle the RAM dump around enough as it reads it to make it compatible with the newer office version, or interpret what is in this RAM dump so it can load it into the newer office. I suspect since OpenOffice would have NEVER had the option to load the RAM dump back in straight they probably ended up with a more robust .doc interpreter.
OOXML ("Open" Office XML) was kind of a greasy ploy against open office and the ODF (Open Document Format), but also legitimately gave Office an actual file format rather than some nasty RAM dumps.
I have just installed 6.2 as an upgrade to 6.1 on a Windows system.
All the variations of the Liberation fonts (mono, sans, serif) looked like mono. Same with the Deja Vu fonts. The Microsoft fonts (Times, Arial, etc.) were OK.
It looks like a typical problem with free software: more enthusiasm than accuracy.
Look up "Microsoft Truetype core fonts for the Web" (or variants thereof), and download them.
This will give you Andalé Mono, Arial, Arial Black, Comic Sans, Courier New, Georgia, Impact, Times New Roman, Trebuchet, Verdana and Webdings.
Bask in Microsoft's 2002 fontal generosity!
[You knew you wanted Comic Sans...]
Looks fresher and keeping the option to revert to menus is exactly what the zealots at MS should have done.
I've migrated to Office ribbons a few years ago but still need to occasionally delve around searching for stuff whereas it would be traditionally under a logical menu system. I'd like to move to LibreOffice permanently but there are still stoo many quirks unfortunately for business use and compatibility. Been trying to migrate permanently since it was the German owned StarOffice in the 90's !
That said, I've noticed that with the CSV file downloads from various banks Excel 36* doesn't correctly open / parse them, even when customising the field settings, whereas LibreOffice works flawlessly, so LibreOffice is set to be my CSV default programme. Hope to drop Office for LibreOffice one day...
Probably MS trying to get more people to use XLSX
I don't think so. It's probably more to do with the fact that double-clicking on a CSV and just hoping it will work, often doesn't because of encoding, etc. Meetup laughingly offers attendance lists as Excel files, but actually provides CSVs which Excel chooses the wrong encoding and "helps" with the dates. Importing as data encourages people to check the structure of the file first.
But CSV is in any case my view a poor format for data exchange largely because it's untyped, so passing round XLSX files is generally more reliable (just as long as they don't have dates < 1900-01-01).
Yes, there is excellent support for old or obscure file formats... LibreOffice grew the ability to import mid-90s ClarisWorks files a few versions back, allowing me to view old university lab reports I had backed on an old Zip disk (!) Reading them I was left struggling with the question of whether I actually ever knew the material or was just blagging it because the intervening years had certainly wiped all knowledge from my brain...
Learning regular expressions is indeed pretty hard for ordinary users. (Guilty!) Moreover, most ordinary users probably won't use them often enough to remember any of them. (Guilty!) That's why I consider "Alternative Find & Replace" (aka "AltSearch") to be an ESSENTIAL extension for LibreOffice Writer. It does the REGEX work for you, and it's always the very first extension I install. (To be honest, AltSearch's functionality should really have been integrated into Writer by now.) Maybe, if we're lucky, AltSearch's author will develop a similar extension for LibreOffice Calc's new REGEX functions. If not, it's back to trying to learn and remember regular expressions ... aka "REGEX torture."
Still not as good as the ePub Plugin for 5.x Writer.
Both of which not as good as Sigil or Calibre.
Calibre is easiest and good when you understand the options. On Linux uses DOCX or ODT, on Windows DOCX or DOC (not sure on ODT).
Save as Word *OR* Open DOCX from Writer to Calibre seems to be a bit better than using ODT.
However source/Edits best done in ODT format.
Calibre is best solution to manage eReaders and create ebooks in ePub, mobi (old Kindles) and AZW3 (newer kindles). Or even ancient ebook formats. Pick Tablet output so it doesn't mess your internal image formatting. Download, browse for or create covers in Calibre. Add CSS rules if needed.
However, once you understand how Writer paragraph and page styles work, I'd rate even 5.x LibreOffice Writer better than any version of Word.
Importing existing user dictionaries is easy.
Like all WP, you DO need to change the defaults on Grammar and substitution.
"However, once you understand how Writer paragraph and page styles work,[..]"
That is its major shortcoming. Its handling of styles is way too complicated and has an unnecessarily steep learning curve, and one of the reasons why I do not use it except for converting/accessing (*cough*Word*cough*) files that do not open well in other programs. Word has similar problems re. styles. It's probably not surprising that LibreOffice has them – they're inherited from trying to be similar to Word – but it seems that LibreOffice's devs don't realize they have a problem there, and do not improve it. Sadly.
Compared to Nisus Writer – which uses an extremely simple and elegant way to handle styles transparently and easily, with an extremely flat learning curve – , I wonder what those developers of LibreOffice are thinking.
On the Mac, Libre's other big shortcomings is that it internally handles fonts and referencing to fonts completely different from all other apps. Almost every document I try comes up with the font in italics, and often uses the wrong font. Which means I have to go in and fix the styles, which with the above problems of styles is a pain to work.
@ elgarak1
..."trying to be similar to Word"
There is the elephant in the room that people criticising LibreOffice/OpenOffice seem to ignore/gloss over (not suggesting you are doing it - this is just a handy entry point): I suspect the problem the devs have is that to get traction, people expect it to be feature similiar to MS Office. In order to do that, ALL Office's quirks have to included as well, otherwise its behaviour will not be the same.
I am certain that the LibreOffice/Oo.o devs can think of a miriad ways to improve on MS Office, but have their hands tied behind their backs.
Not being a programmer, I have often wondered whether it would not be possible to have a super suite that has wonderful features that people want, and then have a MS Office compatibility mode when saving, that warns you about features that will not render/work under MS Office.
It would be nice if one of the devs could weigh in on this.
The real absurdity is when the Latex enthusiasts claim you can "just write". Oh no you can't, not when you have to remember the markup for paragraphs, headers, and lists. And tables are hell! The nearest I have come to "just write" is Microsoft One Note, where even tables are a doddle.
I refer to the original desktop One Note, not to the cloudy abortion MS is trying to impose on us.
As someone who has written two physics theses with LaTeX (including numerous tables and lists), and countless science papers with Word, you are wrong.
Give me LaTeX any time. It's not the best way to get things done ... but it's miles above Word or Libre.
If it weren't for maths, i wouldn't use either nowadays, though. I would go with Scrivener or Mellel for any lengthy science or technical writing that does not use much math. Say, like 'soft' sciences like history.
Enthusiasts through the history of computing have claimed that their favourite programme was easy to use and anyone could. I first met the lie in about 1981. When the schools introduced a word processor ( might have been called Wordstar) to staff, with precisely that claim. In reality a skilled admin was needed to do anything above the most basic, since formatting required confusing and complicated mark-up codes to be inserted. And that is not conducive to rapid typing -even assuming the teachers would or could learn all the fucking codes. Even the admins kept a list taped to their desks.
Oh good.
Ribbon: Garbage
Fisher Price version XP: Horrid, much easier on eyes switched to Win98/Win2K/NT4.0 style
Vista Aero & Transparency: TERRIBLE. Though you can turn it off and make it like Win98/Win2K/NT4.0/XP style
Win7: Stupid flat creeping in, basically a Vista Service Pack
NT4.0 -> Win2K -> XP ->Vista/Win7 -> Win8 -> Win10: Increasingly inconsistent GUI, and Explorer behaviour and splintering of where things found.
Switched permanently to Linux Mint + Mate Desktop (effects off) TraditionalOK theme about 2 years
Switched permanently to LibreOffice about 3 years ago (I'd used Star office and Open Office occasionally, but mostly MS Word & MS Excel since 1993)
So now left the insanity. Wiping remaining Win10 + MS Office shortly.
So called Modern GUI design is GARBAGE simulation of monochrome laser printed paper designed by Graphic designers that seem to know NOTHING about GUI research from 1970s to 1990s. It's gone downhill since 2003 with either insanely flat monochrome or drug addled hyper realistic 3D animations, AKA eyecandy.
The GUI is NOT a piece of art or a Graphic Design statement. It's has to be simple, consistent, work with mouse & keyboard (Stylus & Touch are specific use areas, not a replacement for text / figure / navigation creation). It must not distract or require study.
Oh, and also an application should NEVER force its own GUI unless a simple widget. It should use what ever paradigm / styles / appearance etc the desktop/GUI is set to. You can even write Java that does that instead of a designer selected (or Sun/Oracle default) madness.
Do you hear me Mozilla? Why should I have to install plug-ins to have a sane GUI, or switch to Waterfox because you want to make Thunderbird & Firefox look like an escaped madness from Google or a tablet/phone?
ack - I *LIKE* the "last decade" UI!
double-ack on how foul ribbon and 2D FLATTY McFLATFACE FLATSO "creeping in"
I want a REAL menu, a REAL toolbar [that I can configure and remove if I want], and *NO* *FORNICATING* *FAT-FINGER-FRIENDLY HAMBURGER*
In other words, "last decade" for me.
If it worked, *WHY* *CHANGE* (and force EVERYONE to change?)
Why do I love KDE Plasma so much?
Well, many reasons but one of the biggees is that I make the layout and even the workflow suit me rather than vice versa. Toolbars configurable, apps configurable, kickstart menu - comes in up to five different forms - configurable, etc, etc.
Add in unbeatable custom keyboard shortcut and... wow!
> "So called Modern GUI design is GARBAGE simulation of monochrome laser printed paper designed by Graphic designers that seem to know NOTHING about GUI research from 1970s to 1990s."
Now that's just not true. Today's graphics designers have to have a thorough understanding of what makes a good interface; otherwise they risk making one.
Hah! Pyramids in Egypt are more than 2000 year old and some still find them beautiful.
They were probably much more beautiful when built, but apparently someone thought they looked too -- I don't know, pharaonic? -- and removed the limestone cladding.
So much for progress ...
Happens all over, walk around St Andrews at the end of Fife here in Scotland. Note the ruined Abbey and Cathedral at the end of the main streets. Now wander about the older parts of town looking at building walls and you will notice odd stones in them.
After the Reformation and the dissolution of the monasteries (which happened in Independent Scotland just like in England (and Wales) the good burghers of St Andrews used the good, well dressed masonry of the ecclesiastical buildings as a quarry.
Good Scottish thriftiness you might say, but it happened all over, why most of such sights are thorough ruins. It's just in St Andrews the Medieval town is still largely there and in close proximity to the 'quarry'.
BTW the Castle is still rather more intact and well worth a visit, it has a still intact mine and counter mine from a siege and you can go down it with modern hand rails and electric lighting to help with the claustrophobia.
It's not just the secular sector recycling stuff from monasteries. St Albans Cathedral has lots of bricks whose distinctive size indicates that they originated in the Roman wall.
I read somewhere that until the 19th century, over half the cost of most buildings was incurred transporting materials to the site. So the temptation to re-use anything that's lying around is understandable.
Newgrange, Stonehenge, Sumerian & Akkadian art.
“Epic of Gilgamesh" about 5,000 years old
The wheel.
See Picasso's comments on cave paintings.
Codex: UI to information using PAGES, about 2000 years old. Scrolls suck. Why don't browsers have a paginated mode like web on ancient Kindles have?
Codex (pages) allows random access. Scroll is only sequential.
Bring back absolute position knobs for volume and a knob + scale tuning, Quite possible with a grey encoded optical disc for volume and LCD tuning scale as on some Android phones. A few momentary action buttons and a menu that needs the manual is a RUBBISH user interface for car radios or portable radios.
Physical page buttons, menu, etc on eReaders, tablets, phones (camera button, zoom etc). Touch is great for links or highlights. Rubbish if used for everything. I blame Ives and Jobs.
"The new version 6.2 is the first release in 2019. 6.0 appeared just over a year ago and 6.1 turned up in August."
I wonder if it's time to give 6.1.5 a spin. Last time I tried 6.1 it seemed problematical in various respects so I stepped back to 6.0. Unreasonable? No, at that time 6.1 was the bleeding edge just as 6.2 is now. 6.0 was the choice for "just works".
Why not take the AppImage(s) for a spin? There are AppImages available for Still (6.1.5), Fresh (6.2.0), and Alpha (6.3.x). If you are nostalgic about 2014, there are AppImages for older versions going back to 4.3.7.2.
I have LO installed on my Mint system, but I have used the AppImage for day-to-day work for nearly two years. I have not had any problems running the Still and Fresh editions side-by-side, and the AppImages have never messed with the installed version or the flatpak (spit).
http://libreoffice.soluzioniopen.com/index.php/stable-2/
Download the AppImage, make it executable, and run. If you do not like it, delete the file and it is gone.
"I wonder if it's time to give 6.1.5 a spin."
Seems OK. The bundled icons seem to be the extremes of line drawings my granddaughter would have discarded when she was 7 and the slightly garish. Fortunately the Oxygen add-on still works. I don't know why, under Linux, they can't just use the current icon set from the OS.
I wonder how long it will take 6.2 to settle down. The prospect of being able to use KDE file dialogs on KDE is appealing.
well if all 32-bit support goes away, a lot of OLD computers could still be supported by individual distros, and of course 'build from source'.
On FreeBSD, 'build fro source' was pretty much the only option for quite a while. As I recall it takes the better part of a day on a 2.4Ghz quad core.
It's really not a problem so long as the code does not ASSUME 64-bit, at which point it would have to be patched...
(why 32-bit? it runs a tad bit faster, uses a bit less RAM, and is more compatible with older hardware. What's the harm of supporting it?)
A few modern ARM chips aside-- plenty of 32-bit chips have terrible performance per watt, and of course your energy use includes the mandatory aircon during several months. I actually do like being able to run old stuff and not have to drop cash on "nice" things-- still running the Athlon II M300 craptop is a fine example-- but it's just not worth it over time. Right this minute it contributes to Operation Survive The <expletive deleted> Ohio Winter by spewing heat, but it's still the most useful form of energy, and the most expensive kind of heat, and the biggest waste, and I don't like it.
Suppose that all you care about was your F@H score-- you intend to run some ballpark figure all the time no matter what. Which CPUs and GPUs are going to cost the least --up front? --over the course of two years?
And why does EVERYTHING run faster and use less RAM? It can't always be true but it always gets trotted out and I'm tired of it. I remember when an old GCC took forever to compile something like KDE because it had started using swap space-- before GCC switched over to its C++ implementation, that never happened. Yet everyone "knows" C++ is what you do in order to be more faster and more better. It's really irritating.
Even VIA has had all their processors be 64-bit for over a decade at this point. If you still have ancient devices that don't support it, odds are they're too slow to run modern software that would take advantage of 64-bit in the source code anyway.
But after posting that I thought "Am I sure you can mail merge directly from a spreadsheet?"
So, after googling for one long minute I found this article from 2011 that explains how to do it.
Ha! yes, then actually try it.
It might have worked in 2011 when the article was written but now it's busted.
Create a Writer doc, use "Exchange Databases" (wtf?) as described. Add some fields to the doc.
Save.
Now open it again.
The link to the spreadheet is gone.
Any attempt to perform the merge crashes Writer.
Thanks for that!
I didn't realise you could do that. Now I can edit the source data without using the lame Base UI for it.
It now points out what the problem is - the merge only runs from "registered" databases. he temporary link to the speadsheet isn't registered.
This means that
1) You end up with your global database space being full of all the mail merge sources you have ever used, and
2) you can't ship the source doc and the spreadsheet to someone else for them to do the merge.
So, still massively broken.
and why it is and will remain POS together with all applications for it
[root@linux program]# cat /etc/redhat-release
Red Hat Enterprise Linux Server release 6.9 (Santiago)
[root@linux program]# ./soffice
/opt/libreoffice6.2/program/oosplash: /lib64/libc.so.6: version `GLIBC_2.14' not found (required by /opt/libreoffice6.2/program/oosplash)
/opt/libreoffice6.2/program/oosplash: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.18' not found (required by /opt/libreoffice6.2/program/libuno_sal.so.3)
/opt/libreoffice6.2/program/oosplash: /lib64/libc.so.6: version `GLIBC_2.14' not found (required by /opt/libreoffice6.2/program/libuno_sal.so.3)
And no, I'm absolutely NOT planning to upgrade (wipe and install new as there's no damn upgrade from RHEL6 to 7) my OS to just run some application.
But that's probably how some in the FOSS world like it. ® indeed
This post has been deleted by its author
So you're upset that installing the latest version of a third-party desktop program onto an 8-year-old version of a server OS doesn't work? RHEL 6 ships with LibreOffice 4 anyway, so why not just use that until you're ready to finally jump to RHEL 7 (or 8 judging by how long you've clung onto RHEL 6 for)?
You really should be in the planning stages now to migrate to a later R\HEL - you could always try CentOS 7 or RHEL 8 beta in a VM to dip your toes in the water. There's less than 2 years of RHEL 6 support left and more and more third party software won't work with it. I kept Chrome going on CentOS 6 for 3 years after Google "broke" it on that platform, but had to throw the towel in when they switched to GTK+3,
I'm upset because in year 2019 Linux software developers can't grasp a basic idea that they need compile software in the way that makes it compatible with as many operating systems as possible and not only for systems they use themselves or for systems they think others should be using.
You know, how they do it for Windows, same LibreOffice works perfectly fine on older than RHEL6 Windows 7.
Should I also mention that System Requirements page for Linux of LO version explicitly states "glibc2 version 2.5 or higher" while in reality it's 2.14 or higher ? Says a lot about quality.
> I'm upset because in year 2019 Linux software developers can't grasp a basic idea that they need compile
Oh I see, we're all so stupid. Luckily for you we know how incompetent we are so we make the source code available so that you can compile it yourself.
Don't forget to share the .spec file when you're finished, thank you.
"I'm upset because in year 2019 Linux software developers can't grasp a basic idea that they need compile software in the way that makes it compatible with as many operating systems as possible and not only for systems they use themselves or for systems they think others should be using."
Surely this is less to do with Linux developers in general and more specifically to do with whoever manages the binary packages for your choice of OS?
Firstly, binary packages are on LO site and they can be installed by a user without requiring root access.
Secondly, it's a very common issue that affects pretty much all recent software - Chrome, Opera, lots of other stuff that is compiled only for particular version of OS or set of libraries, the one developers use themselves. In many cases there's no source code and there's no excuse to not provide statically linked binaries, amount of spare CPU power, memory and SSDs on modern computers totally cover for minor performance impact and larger binaries.
"Firstly, binary packages are on LO site and they can be installed by a user without requiring root access."
Oh, so they are! What "OS" is "deb" or "rpm"? It seems a little odd that any project would provide binaries without specifying exactly what they are built against instead of just directing users to their specific OS repository. I wonder if those downloads are the reason some people had issues installing it?
> What "OS" is "deb" or "rpm"?
There are two main packaging systems. These are used by the two main hierarchies of Linux distributions. Debian based systems (Ubuntu, Mint, ...) use .deb (debian), Red Hat based systems (Fedoda, CentOS, ...) use .rpm (redhat package manager).
> It seems a little odd that any project would provide binaries without specifying exactly what they are built against
If you are running Linux and don't know whether you are using .deb or .rpm then you should stick to the repository.
> instead of just directing users to their specific OS repository.
The repositories are usually a version or two behind the bleeding edge.
Essentially anyone who has been running RHEL6 in recent years should be well aware that its libraries are ancient. I'm certainly painfully aware of that fact (we've had an upgrade to RHEL7 on our very long to-do list for some time. That's partly the point of it.
The rest is just a noisy rant, the whole point of FOSS and distros is you use the software built for your system. Anyone with good reason to be still running RHEL6 should be more than capable of building their own software for it. The real pain comes when library versions have changed sufficiently to make even that difficult, which is at heart the reason recent software will be packaged for recent libraries.
> Linux desktop and why it is and will remain POS together with all applications for it
If your distro is too old for the normal binaries, or to use Flatpak (the better solution for any distro released in the last 4 years), use the AppImage:
https://www.libreoffice.org/download/appimage/
"LibreOffice AppImages were tested and run on these old versions of stable distributions and their newer versions - Ubuntu 10.04 LTS, Debian 6, Fedora 12, openSUSE 11.3, Mageia 2, PCLinuxOS 2013.07.15, Puppy 6, Salix 13.37 (Slackware), CentOS 6"
Libre/Open's equation editor is outstanding. I'm far more productive in it than MS's, which pretty much requires constantly moving a hand between keyboard and rodent. Not as powerful as LaTeX, but not everything is a thesis.
Second thing is that calc can ingest truly enormous amounts of csv data and still run very fast. Ive thrown hundreds of MB into it, with total confidence. Excel? segfault.
Great work!
First of all, it ought to support all input methods that users of that platform are familiar with.
The fact that Mac users have to forego the very easy way in which accented characters (as in "those directly accessible from a layout, for example éëêè on US/UK keyboards) can be entered on a Mac is stupid - as a matter of fact, I don't know if it's patented but if not, I'd love to see that approach in LO as what is presently there is painful, for all platforms. For Mac users, the solution tends to be installing NeoOffice instead - worth the few euros per year they ask in support IMHO - if you work in multiple languages from a US/UK layout you have earned that back in time savings in no time. I'd also love to see this approach in Linux - it's simply more practical as you can do it straight from the keyboard, with no mouse involved.
Secondly, the installer really needs work if LO is ever to make it onto end user machines worldwide. It needs to be ONE installer, and it needs to work in the language of the user from the start. To install LO, you need to install the (American) English version, then "patch" the UI into the user's language by means of a language pack, then start LO (still in American) and set it to your preferred language, then restart. Updating: ditto. This is despite the fact that someone installing a language pack is very likely to want to set it to that language, so why not ask and set it? It needs one, multilingual installer. Maybe make the installer load a language pack on startup and then run both installer and UI setup in one go, but at present it's a major barrier, also for volume deployment.
One benefit not mentioned is one of speed: LO has endeavoured to introduce spreadsheet code that supports parallel processing (i.e. using multiple threads) of cells with similar formulas. As far as I know, that support went public in the 6.2 release, and I'd welcome news from anyone who has tried it (my spreadsheets are not that complex that I'd notice a change in speed :) ).
I really *want* LO to be wider successful, but the above are serious barriers as especially the installer is the first thing a new user ever sees of a product. Asking users to change their approach to input is also not helping, but comes IMHO on a second place.
Writer fails on a the most basic task of automatic page numbering, something that can be done in every other word processor in existence with literally two clicks.
Let me repeat - word processor that has gazillion options after 10+ years of development does not provide a simple and straightforward way to insert page numbers starting from a second page (2 on second page, none on first) without messing with styles.
For real.
> does not provide a simple and straightforward way to insert page numbers starting from a second page
I do that on pretty much every report that I write¹. It takes three clicks exactly (or you can use the keyboard).
Respectfully, I suspect that your problem is quite simply that you do not know how to use a computer². Users who think they're clever tend to be the worst. :-(
¹ I obviously only did the layout once and saved it as a template years ago.
² If I have to use another word processor than the one I'm used to, I do feel pretty lost, but it wouldn't occur to me that that's the other software's fault.
Are you smart enough to understand "without using styles" or all you know is how to use a computer ?
I think he/she is smart enough to understand that styles are the single most productive thing to learn when you're allowed near a wordprocessor. It's the equivalent of learning how to drive a car and not using it in first gear only.
Writer fails on a the most basic task of automatic page numbering, something that can be done in every other word processor in existence with literally two clicks.
I've been only been using LibreOffice since it started as StarOffice on OS/2 (and I paid for that) so I may have missed something, but I have been *always* been able to do that.
Let me repeat - word processor that has gazillion options after 10+ years of development does not provide a simple and straightforward way to insert page numbers starting from a second page (2 on second page, none on first) without messing with styles.
Aaaaah, now I get it. You're one of those people that formats everything by hand because you have no idea just how much time it saves to use styles for it. Yes, you need to apply a "first page" style to your, well, first page because it's been designed explicitly for that purpose.
You really, really need to get to grips with that. Personally, I think educating people about styles is the single most beneficial thing you can do if you want people to be more efficient and faster. Screw everything else, if you learn to separate content, structure and styling you will be more effective on any platform, in any word processor.
And you will no longer display ignorance by blaming your tools.
"Aaaaah, now I get it. "
No, you obviously don't. How you are formatting your documents is absolutely irrelevant to anyone else, you can do it through your butthole swinging on a rope, noone cares.
The point is that MS Office, where most users coming from, has that option in a menu and LibreOffice does not.
How you are formatting your documents is absolutely irrelevant to anyone else
Yes, if you only write love notes to yourself. In any other environment you write it to exchange information, and to do that correctly, the approach is structure, content, then formatting. That gets you out of that rut that a document doesn't *look* the same on another machine or system, a ploy that Microsoft has been using for years just to get people to upgrade. It means you can exchange information with, gasp, other people, people that use, gasp. OTHER operating systems and even other word processors.
Word processors, for instance, that don't have a productivity destroying ribbon and hide all functions in entirely new places so they can sell another upgrade..
The major reason that documents look different on random machines is the fonts that are installed - or not.
Never been an issue in LO. Choose "Properties, Font", and make sure you have "Embed fonts in document" and "Only embed fonts that are used in the document" ticked.
You're right. Inserting a simple page number into the footer for pages 2 et seq,, leaving the first-page footer blank, requires a minimum of 9 clicks and 2 different menu sequences, and the procedure doesn't replay correctly as a straight recorded macro (at least not without editing). Basically, you have to format the page to turn the footer on and use a different footer on the first page, and then you have to go into the second-page footer and insert a page number (and optionally format it). That's only one of the reasons I use templates with well-designed styles for most of my work in Writer. But yes, it's too confusing for new users, and it's too tedious for any user. The developers should probably add a special-purpose command or macro for this common function.
I am not quite sure I understand the first problem.
I write English, Danish (æ, ø, å), German (ü, ö, ä) and Portuguese on a Portugues keyboard MacBook Air, and all the accents on the KB as well as the ones reached by "alt" works just the same in LO as outside.
I write English, Danish (æ, ø, å), German (ü, ö, ä) and Portuguese on a Portugues keyboard MacBook Air, and all the accents on the KB as well as the ones reached by "alt" works just the same in LO as outside.
Now try that if you have an English keyboard...
Interesting. So to rephrase my understanding, the key combinations that work in OS X generally do not get through to LO if you have an English keyboard, but they do if you have a non-English?
I do have an interest in this as I just could consider buying a US/UK 'book.
Found this, btw: https://sites.psu.edu/symbolcodes/languages/europe/nordic/#macaccent
Try activating the MacSO built-in key modifier. In any other application, holding down "e" for instance, pops up a menu so entering éëêèęė and ē are easy, and not all of these have direct alt shortcuts.
It's a simple system that works across all languages and MacOS applications. Except OO and LO.
Ever since I bought my new laptop a month back I've been running LibreOffice. It seems work very well.
If I need to do some statistical analysis that LibreOffice can't handle I can still just export to csv and use R or Python to do it.
In fact something like ggplot generally solve the template problems as far as my graphs are concerned.
There are plugins for Zotero that work perfectly fine for citations.
And I can also still write macro scripts if I need to.
Much better than buying a subscription to Office 2019.
This updated interface is good, I have a choice between Menu, Ribbon or both together.
Seriously MS Office is giant waste of money lol!
When OO made its move from Star Office to Open Source, there was a nice, comment filled load of beta testing.
I remember someone from the US being very unhappy that it was not following their "standards" in date formats, measurements etc. Apparently, it let them change to their preferences but it didn't default to assuming Letter was a standard paper size far superior to A4 because it used units of measurement Henry VIII would have been familiar with...
The only bit they looked like they were getting wrong was they were not interested in spelling "metre" correctly. Have they fixed that?
Speaking of paper sizes, I remember the first released version of OpenOffice could not actually print (the relevant module probably contained some 3.party code that could not be open-sourced, and needed a rewrite). Of course, that was fixed soon, but it was a case or "release early" taken a bit too far.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
