back to article LibreOffice slips out another 7.0 beta: Spreadsheets close gap with Excel while macOS users treated to new icons

The Document Foundation has released a second beta of LibreOffice 7.0, with general availability expected at the beginning of August. The addition of soft edges support fixes some cases where a PowerPoint graphic looked completely wrong in LibreOffice. The addition of soft edges support fixes some cases where a PowerPoint …

  1. Charlie Clark Silver badge

    Compatability

    which is now in the 2013 version rather than the 2007 version. This may cause problems for Word 2010 users, to whom the advice is to "upgrade to LibreOffice".

    Unlikely: the transitional spec was really developed for MS Office 2007 and was incomplete, so that changes were made even after it shipped. Microsoft itself deprecated Office 2007 a while back so it's not uncommon to come across OOXML documents that Office 2007 cannot process completely but I've not seen such problems with Office 2010 which saw far greater adoption, largely on the promise of stable document formats.

    1. Anonymous Coward
      Anonymous Coward

      Re: Compatability

      Not unlikely at all. I've sent someone running Office 2010 a "Strict Open XML Spreadsheet" as saved by Excel 2013, and he could not open it without installing Microsoft's OOXML Strict Converter for Office 2010. Even Office 2013 and perhaps later versions still default to OOXML Transitional.

      1. Anonymous Coward
        Anonymous Coward

        Re: Compatibility

        I'm hoping that the usual is true: for many years, compatibility problems between various variants of Office could always be easily resolved by opening them in LibreOffice or OpenOffice and then saving/exporting them in the new format.

        As a matter of fact, very large Word documents would normally collapse under the weight of broken formatting codes generated by the idiocy of including them badly in cut & paste operations and then refuse to open in Word. The simple trick there was to open them in OpenOffice or LibreOffice and then save them again, as OO and LO would just silently clean up the mess and then spit out something that Word itself could understand again.

        After that you could normally clean up the various other formatting problems.

        That all said, 99% of the problem comes from people not having the slightest clue of how to use document styles which results in a pile of inappropriate formatting dwelling in documents, just waiting for the slightest bit of bit rot to take revenge on the unsuspecting user..

    2. Len Silver badge
      Devil

      Re: Compatability

      If it is OOXML compatibility you're after just stay away from Microsoft Office and you'll probably be fine.

  2. Zippy´s Sausage Factory
    Thumb Up

    I'm a big fan of LibreOffice so it's nice to see a release that seems to be strengthening it. I already use LO a lot at work, as it tends to crash much less than the Micros~1 version (although that might be a consequence of our support guys insisting we stick to the 32 bit version, presumably forever)

    1. Peter2 Silver badge

      We have to stick to the 32 bit version because if we don't the program that everybody uses daily which generates the letters for them won't work.

      It's also why we aren't using LibreOffice. Doing so an generating everything manually would cost more than the software licenses.

      1. Charlie Clark Silver badge

        While I can understand sticking with 32-bit because of the add-on/integration problem, I don't see what's stopping anyone providing something similar for LibreOffice / OpenOffice which have both provided a rich API for automation for decades.

        1. Captain Scarlet Silver badge

          I tend to find we have add-ins forced upon us for things like specific non standard charts the incoming MD wants. These then spread to other staff and they just stick like an annoying tic.

        2. big_D Silver badge

          Because the providers go where their clients are, not where they can go. And the clients go where the software is...

          It is a vicious circle. We have add-ins for our telephone system and they work with 32-bit MS Office and 32-bit MS Office only. The telephone system manufacturer isn't interested in 64-bit or LibreOffice, because they have a 32-bit client already and most of their customers use 32-bit MS Office, those that don't will miss out on the functionality.

          Same for the DMS system.

    2. Captain Scarlet Silver badge
      Coat

      The majority of add-ins tended to be 32 bit a number of years ago, not sure if its still the same.

      Also an accountant able to use more than 3GB of memory in excel spreadsheet is just plain scary!

      1. Charlie Clark Silver badge

        They certainly tend to stick around once there inhouse. Got a similar problem with a client currently migrating from MS Office 2010 to Office 2019 and sticking with 32-bit because of the fecking add-ons, which are probably no longer even maintained.

        Good that there are now other options with things like PyXLL.

      2. John Brown (no body) Silver badge

        "Also an accountant able to use more than 3GB of memory in excel spreadsheet is just plain scary!"

        Likewise, accountants given access to a spreadsheet capable of a "database" with millions of rows!!

  3. DrXym Silver badge

    Changing the backend is one thing

    One thing I notice on Windows in 6.4 is a serious rendering issues when resizing windows. For example, if I resize the window frame horizontally or vertically the repaint of the content lags badly behind. In the meantime it suffers nasty layout juddering and gaps with black / garbage and/or suffers clipping issues depending on the new size being larger or smaller.

    So changing the background is great and may speed rendering a little but the perceived experience to the user matters too. A little spit and polish on the user experience and usability would a go a long way to making the suite more popular.

  4. Porco Rosso

    Any news of a Libre_mail client

    Its nice to see them improve the suite and accelerate the development.

    a big thank you.

    but is there any news if they will develop/create/integrate an Outlook/Lotus Notes/Thunderbird/Zimbra alternative ?

    I find it a big missing point in the suite.

    1. Martin Gregorie Silver badge

      Re: Any news of a Libre_mail client

      Evolution does everything I need for communicating with people and managing future tasks: e-mail, contact management, calendar plus a task list and memos.

      1. ovation1357

        Re: Any news of a Libre_mail client

        Correct me if I'm wrong but isn't Evolution only for 'Unix-like' operating systems - i.e. not Windows.

        Personally I never got on with it and I'm a huge fan of Thunderbird which is, to me at least, the holy grail of mail clients. Sadly it doesn't have native support for Exchange but I have previously used it with great success using the non-free but very cheap Exquilla plugin.

        I'm currently stuck on Windows 10 with Outlook 2016 at work and I'd jump at the chance to use Evolution instead (I've already tried Thunderbird but with no joy because of GPOs and general lockdown)

    2. Captain Scarlet Silver badge
      Mushroom

      Re: Any news of a Libre_mail client

      Lotus Notes was never about email (Until it was to late), why do you think the email application template was clunky for so long! It was always about having networked applications and hope to god the develop knew how to handle multiple edits of the same record at the same time!

    3. Lee D Silver badge

      Re: Any news of a Libre_mail client

      I would like to know too.

      I'm still running Opera 8 (?) purely to collect and sort email (I moved on to other things for my browser, and the "new" Operas have basically abandoned mail functionality.

      Massive SQLite database underneath it with 20 years worth of email over a dozen accounts, searchable in a trice.

      Tried everything - even Pegasus Mail at one point. Can't find anything that I can get on with or that is as fast when searching that amount of email. An LO alternative to Outlook would be great.

    4. Len Silver badge

      Re: Any news of a Libre_mail client

      When Thunderbird was going through one of its bi-annual “We’re leaving Mozilla” processes there was some talk at the OpenDocument Foundation of taking it under the wings of the Libre packet but I don’t think it went very far.

  5. Anonymous Coward
    Anonymous Coward

    "presuming you believe that such a large spreadsheet has any plausible use"

    Reminds me of a vehicle dynamics project I was working on in the early 2000's.

    I was told to "collect lots of data" so it could be post-processed. The data was in CAN messages, so I just stuffed the lot in the specified CSV format file. Each run fitted easily enough on a CD, but the team that were processing them reported that "they wouldn't open". Turns out they were trying to use Excel to read them, and it didn't like the 2.5 million "rows". Was never a problem using "proper" data analysis software like Matlab...

    1. My other car WAS an IAV Stryker Bronze badge

      Re: "presuming you believe that such a large spreadsheet has any plausible use"

      Funny, I was going to make a similar comment about trying to process CAN logs in Excel. Vehicle tests sure can make data in a hurry, especially the more things on CAN there are*.

      (Yes, I was familiar with MATLAB and should have used it, but I didn't.)

      * Especially if you're saving the entire unfiltered bus, because 90-95% of the content is probably pointless to the problem at hand. Filtering on header content -- especially using SAE J1939 protocol: PGN or source address -- helps a lot. But then again, in case questions arise, you never want to say "oops, I didn't record that one" in case you don't get a chance to repeat the test (which happens a lot).

      1. Nick Ryan Silver badge

        Re: "presuming you believe that such a large spreadsheet has any plausible use"

        Hell yes. Sometimes it's the incidental messages that show up the problem rather than focussing on the specific device or module that is exhibiting problems. Malformed messages on a different device, or an unexpected device ID are usually a sign of the problem being elsewhere.

        I almost enjoyed giving myself a crash course in everything CAN bus related and writing a protocol handler to process the messages on a CAN bus controller. I had to re-write it from scratch because the previous developer treated it as single point to point serial connection rather than a bus therefore only one thing could be caused to happen at once regardless of the number of nodes connected and as for configuring and responding to asynchronous broadcasts? Nah, just he just polled the device when interested. It was amazing that the project got as far as it did. /sigh

  6. Radio Ballet Shoes
    WTF?

    Input of accents on mac

    Have the developers fixed the accents input issue on macs? The one where most of the other apps allow you to hold down a key to scroll through the accents for that key character. Libreoffice doesn't! It seems a really really stupid omission to me.

    1. W.S.Gosset Silver badge
      Headmaster

      Re: Input of accents on mac

      Oh hey! A new Mac OS X feature which is actually useful!

      Of course, built-in since 1984 MacOS 1.0 is the ability to enter almost ANY accented character in 2 keystrokes. So maybe they just assumed people remember that.

      Eg, hold down Option and press "e", then press "e". Voila!

      Option i , i. Option u, u. Option a, a. Etc.

      Fire up KeyCaps (now renamed) and press Option key to see at a glance all the options available. Note Option-Space might be a bit cryptic; it's the non-app-specific nonbreaking space.

      1. W.S.Gosset Silver badge

        Re: Input of accents on mac

        God, I miss MacOS. You just got things DONE.

        I remember a Unix admin legend and LaTEX guru at work labouring over a formula for his thesis for 2 weeks, with many yet to come, and he couldn't get it right. Took him into my uni Mac Plus (I was working 2 jobs during my Masters), handed him the whole lot in less than half an hour.

        "I _could_ have done that in LaTEX, you know...."

        "Yeah. But that quick, that easily?"

        "..."

        This used to be trivial knowledge for everyone. People so un-tech they wouldn't notice the power cord was unplugged and complain the computer had crashed, casually banging out stuff like this, and casually hacking their kernel. Without even realising that that was what they were doing, let alone how awesome it was.

        Your now-ubiquitous Autocorrect, btw, was originally just a casual MacOS kernel hack. Thunder II, IIRC. I can't remember right now whether it was Apple or Microsoft first nicked and reimplemented the idea and issued it for free and built-in (in MS's case, first in MSOffice: the MacOS GUI recreated-ish on Windows), but either way, those poor bastards got rolled.

      2. big_D Silver badge

        Re: Input of accents on mac

        How did it know which accent to use, with just 2 keystrokes? úùüû? éèê? etc.

        1. katrinab Silver badge

          Re: Input of accents on mac

          ⌥e e - é

          ⌥u e - ë

          ⌥I e - ê

          ⌥` e - è

          ⌥n n - ñ

          In windows, you can use the UK extended keyboard to get similar features

          You use Alt-Gr rather than ⌥.

          For ë you would use " rather than u

          For ñ you would use ~ rather than n

          For ê you would use ^ rather than i

          For é you would use ' rather than e. You can also do Alt-Gr e, and that applies to the standard UK layout as well as the extended layout, as there are some English words that require it such as exposé which is not the same as expose and some where it is an alternative correct spelling such as café.

          1. big_D Silver badge

            Re: Input of accents on mac

            That makes sense. By 2 keypresses, I was thinking ⌥e, not ⌥e e. The same as compose keys on DEC PCs and terminals.

          2. Pen-y-gors Silver badge

            Re: Input of accents on mac

            For many years on Windows I've been using a little utility called To Bach, originally developed to cope with adding a ^ to w and y in Welsh (It's called a To Bach, meaning Little Roof, or, more formally acen grom) - Basically AltGr + w gives ŵ , default is AltGr+letter puts a ^ on it - â - but it was extended to do more, much like the extended keyboard noted above AltGr+", i gives ï, AltGr+/,e gives é

            A really neat bit of freeware!

          3. Dan 55 Silver badge

            Re: Input of accents on mac

            To be honest on a UK keyboard the accent deadkeys are pretty nonsensical on both Mac and Windows. I ended up using Ukelele and Keyboard Layout Creator to put the deadkeys on the accent keys so it's more like a compose key.

            The popup you get on later macOSes when you long-press a vowel is pretty useful for finding an accent you don't use much.

            1. W.S.Gosset Silver badge
              Unhappy

              Re: Input of accents on mac

              In MacOS it was apparently keyboard-layout agnostic. Personally tested: Australian, US, UK, Austrian/German.

              God alone knows what Mac OS X is doing.

          4. Len Silver badge
            Happy

            Re: Input of accents on mac

            I used to have a colleague with an ö in his surname and I fairly quickly learned the key combination to write his name. Within a week you don't even think about it. Twenty years later and I still know that it was ALT 148 (it was on Windows in those days).

      3. Dave559 Bronze badge

        Re: Input of accents on mac

        The way MacOS makes accents and special character input easier is quite nice, although some of the keystrokes are rather unobvious and obscure, and there are some commonly used characters that it doesn't seem possible to type directly in plain text (eg, ², ³). I am personally more comfortable with the compose key idea which is common in Linux: does anyone know if there is a way to do this in current versions of MacOS? (I did try to look this up a while ago, and it seemed that there were a couple of possible solutions, but they seemed to be somewhat MacOS-release-specific, and I ran out of time to hop around the rabbit burrow to find out where the path eventually led…)

        But it is particularly ridiculous that there really is no easy way to enter special characters on Windows, however (alt+number isn't particularly friendly)!

        1. katrinab Silver badge

          Re: Input of accents on mac

          Easiest way is:

          Cmd - Space

          Type 2, 3, or whatever in the search box

          Select the one you want.

  7. Brewster's Angle Grinder Silver badge

    What do you call a spreadsheet with 1 million rows? A database.

    1. Charlie Clark Silver badge

      Not really, a table at best. But at least Excel provides some kind of data typing, which is why it's become so ubiquitous. I've recently seen this confirmed in a project which relies on data passing through several hands before it can finally get into the database. Using Excel as the file format has led to far fewer errors than I would ever have imagined.

      1. W.S.Gosset Silver badge

        Low-level data format tip: bin CSV, go with TSV.

        1. Charlie Clark Silver badge

          In my experience, CSV/TSV all fail as soon as you have non-ascii data.

          1. W.S.Gosset Silver badge

            In my experience (>30yrs of what is now called bigdata), I have never even heard of such a thing nor am I able to comprehend how it's possible unless you're using some mad manky v.old tool. It's certainly got precisely 0 to do with how you delimit or encode data fields. Non-ascii is non-ascii no matter what you wrap around it, and no matter whether you write the payload _en claire_, base64, or raw binary.

            If your destination can't cope with non-ascii, that's not a problem with the data file's format.

        2. Pascal Monett Silver badge

          Ah CSV. The magic of having addresses with commas in them. I cannot count the number of hair-pulling times I have had to export data into a CSV file and send it off, only to have the recipient call me and say MY file wasn't working.

          Buddy, I exported YOUR data. I you can't be arsed to not bung commas everywhere they shouldn't be, you don't get to blame ME for it.

          1. Alumoi

            I you can't be arsed to not bung commas everywhere they shouldn't be, you don't get to blame ME for it.

            Funny, over here (Romania), addresses HAVE commas as standard: city, street name house number, building, floor number, apartment number,

            Export this!

            1. Anonymous Coward
              Anonymous Coward

              Commas in addresses

              Yep, UK too in at least some instances. Then there's the issues of commas also being used as thousand seperators, so I too prefer TSV to CSV.

              And don't get me started on the XML guys deciding "&" was a special char when it is in thousands of UK company names :-)

          2. Tim99 Silver badge

            I use SQLite to "tidy" recalcitrant data: Command Line Shell For SQLite

            Apparently the maximum database size is 140,000 gigabytes - I wouldn't know, I've only tried it with a few million rows with a hundred or so fields :-)

          3. Pen-y-gors Silver badge

            CSV?

            Ah CSV again. Not so much commas, as the field they're in can be delimited by " in most sensible exports. Real pain in the posterior is how Excel exports data with carriage returns in a cell. Which creates a new line in CSV, which causes some grief when trying to import into PHP - anyone got a suggestion?

            1. Dan 55 Silver badge

              Re: CSV?

              Count the cells in the line, if it's less than expected then take the linefeed to be a continuation of the previous line.

              1. Anonymous Coward
                Anonymous Coward

                Re: CSV?

                Newlines in quote delimited fields in CSV are perfectly valid. So the solution is to get a better CSV parser in PHP which understands this.

                1. W.S.Gosset Silver badge

                  Re: CSV?

                  Spot on.

                  This is just crap coding by the receiver library. To a degree which makes my gut hurt.

                  More-general point: MSExcel (rather unusually) exports CSV in correct full CSQDV form, but do also note my post below. Which is useful for anyone seeking to create their own CSV/text export file.

                  1. Dan 55 Silver badge

                    Re: CSV?

                    Then import from Excel localised to a European language and watch it blow up. Decimal points become commas and commas become semicolons, but the extension is the same.

                    The poster who said throw it all out and use TSVs was right.

            2. This post has been deleted by its author

          4. W.S.Gosset Silver badge

            CSV=poor understanding --> CSQDV correct

            Ah, that's because you don't know how to do CSV correctly.

            See, the _correct_ name for the format is Comma-Separated Quote-Delimited Values, wherein all strings have to be wrapped in double-quotes. Mandatorily. "CSV" is just a shorthand term.

            From what you've described, you simply set IFS to comma and created a file. Well, yeah: broken.

            TSV eliminates this problem since essentially zero text fields embed a Tab character.*

            Just set IFS=^i, print (as we used to say re export to FS)**, job done.

            Less work, better result. Parsimony. Hallmark of a good coder.

            .

            * In the extraordinary event that you have an embedded tab, the format provides for, but unlike CSV does NOT require, quote-delimitation of text strings.

            Please note that any success you've ever had with non-quotedelimited CSV is purely the destination system being forgiving -- you have sent a broken protocol file and it rightly should reject it.

            ** interestingly, the financial markets just in the last year or so have adopted the same term for the same purpose/reason: any explicit issue of data. So analysts and economists now routinely talk about, eg, the last print of unemployment data or the last print of a company's results.

            1. Dan 55 Silver badge

              Re: CSV=poor understanding --> CSQDV correct

              Excel itself didn't put string fields in quotes until later versions. If a string field is surrounded by quotes then an escaped quote is two quote characters (""), otherwise a quote is just a single quote character. So that makes decoding more fun.

              1. W.S.Gosset Silver badge

                Re: CSV=poor understanding --> CSQDV correct

                That's actually part of the spec.

                And also standard across a wide range of other contexts.

                1. Dan 55 Silver badge

                  Re: CSV=poor understanding --> CSQDV correct

                  So you can't really rely on string fields being quote delimited. If you code for a specific version of Excel then all you can really say is you can exchange data with that specific version of Excel.

                  Double quotes remind me of the 1980s, a backslash for escaping characters is more usual now.

      2. Brewster's Angle Grinder Silver badge

        "Not really, a table at best."

        I was being facetious. But multiple sheets?

        "Using Excel as the file format has led to far fewer errors than I would ever have imagined."

        I entirely believe that. I think it bolsters my facetious point: it is a different front end on a database.

        1. W.S.Gosset Silver badge
          Thumb Up

          No, B.A.G., you are entirely correct, and valid. BUT only looking at one part of it (cf 4 blind men and the elephant). I speak as someone who's used and stretched MSExcel since vn 1.04a. Who remembers and uses explicit intersections, for example.

          I describe Excel to people as a range-based matrix engine with an embedded 2D database (with ISAM-style 3GL-primitives linking between data files (intra-db + external)) with an integrated presentation layer.

          99.9% of people mistake the presentation layer for the spreadsheet (cf. Libre Office) and miss all the power.

          And most people only ever use one tiny aspect or another of its full scope. No problem with that. It's not a dick waving exercise; it's a tool, intended to be used any way you bloody like to solve YOUR current problem.

          If I buy a Lamborghini and then use it to pop down the shops to grab some milk, or to tow a trailer of junk to the tip, or to stand on to clean out the guttering, the Lamborghini doesn't suddenly become a shit car. I'm just not using its full potential; I'm just using it for my current micro problem. It's still an awesome piece of kit.

          ----

          Having said that:

          The depth and integration of the core maths' and matrix engine's intelligence of design, and even of how the dependency graph's lazy compilation interprets cells' content (they're converted to raw text then built-up till the last moment then passed to eval())(have a look at INDIRECT() to twig it's a user-accessible EVAL. And yes that does in fact mean everything it might) is truly and never-endingly awesome and inspiring.

          Put it this way. I replaced a custom program in a fund manager to calculate portfolio risk, and an equivalent spreadsheet using multiple whole sheets, with a single formula. Simple matrix multiplication of a covariance matrix and a weights vector. = weights-range (click, drag) * covariance-matrix (ditto) * transpose weights-range (ditto again), and hold down control shift when hitting return to have it treated low-level differently as an Array Formula. Job done.

          Name the Ranges and you're now also Robust to future changes.

          1. W.S.Gosset Silver badge

            Low-level Excel tute

            Actually, everyone should know about at least that tiny but powerful aspect:

            To get a glimmering of the power under-the-hood:

            Create 2 Cols of numbers (eg A+B) where you want to multiply each pair together then add up the total of those per-tuple products. Yes, roughly just a simple Average (sum for i of weight times value).

            Do =SUM(A:A * B:B). Note the result is borken.

            Now do =SUM(A:A * B:B) (identical , ie) but this time hold down control+shift when you hit return to tell Excel to treat this as an array formula. You'll see it denoted by having {} shown around it.

            Note the result is now perfect.

            And yes this example is trivial. The point is that MSExcel is startlingly atomic at its core and absolutely cleanly so, like original unix's core tools. And just as you can build up indefinitely complex pipes and commands and scripts in shell using its clean primitives and It all Just Works, so too Excel.

            1. W.S.Gosset Silver badge

              Daily Datafile processing, Re: Low-level Excel tute

              Now, Grasshopper, consider that Excel's engine has no concept of Cells, only of Arrays ("Ranges")(a raw Cell is treated in calc as a 1:1 Array), then consider that you have only ever used its Default For Donkeys auto facility for lazy people: Implicit Intersection.

              Explicit Intersection is very powerful.

              The syntax is "[Range] [Range]". Ie, 2 ranges (which encompasses 2D array, 1D vector, and 0D Cell) separated by a space. (Put it/use it anywhere where you would currently use what you would call a cell-reference. The engine evaluates and collapses all such references to a single array before passing on.)

              Have a play with that. Notice a lot of what people proxy via INDEX() goes away.

              Then consider that built-in since version 1 is refs to other spreadsheets and files. The full syntax (IIRC -- no excel here to check) for EVERY cell reference is "[filename]sheetname!cellref". In fine relational tuple style, this is stored with EVERY cellref, albeit for diskspace parsimony via pointers (re-evaluated at calc-time). So if you routinely have data files coming in which you need to do Excel mungling on, you don't need to import them/copy them into a copy of the template workbook first. Which is A/ a ballache; B/ introduces serious risk of tired/fuckedoff user error; C/ creates 2 copies of the same data with all the concomitant risks of eg downstream undetected modification of "source" data, accidental or otherwise; D/ is just ick.

              You can simply create an always-external workbook which looks externally for the datafile by name. And can then maintenance-simply embed exactly, only, and no more than the desired downstream processing.

              For bonus points, consider that you could replace the formulae's references via INDIRECT () to simply lookup a single "Please Input Datafile's Name Here:" field. Type a new filename, see the new processing's results. SaveAs (for audit purposes), done.

          2. W.S.Gosset Silver badge

            Re: ISAM-style

            Nowadays that would be "NoSQL". Sorry.

            Modern syndrome: endless re-inventions of "discoveries" and "innovations" which were situation-dependent bodges 70yrs ago, acknowledged as same and rejected for a purer & better technique as soon as technically feasible. By people who actually understood.

            I wouldn't mind so much being called old, if only the kids didn't keep dragging our old discarded clothes out of the attic and dancing around in them screaming that this is and THEY ARE on the ultimate bleeding edge of fashion.

    2. Len Silver badge
      Happy

      I don't think many people will use Calc (or Excel) to create 1 million row spreadsheets but as the output of some automated process they're probably not uncommon. It's nice if your spreadsheet application can at least open them so you can use it for advanced logic to strip all the unnecessary bits.

      Unless, of course, you prefer to write your own parser and regex to strip out all the rows that have values below 0.00153 except when the another column lists "Belo Horizonte", "Carlyle" or "Turku" but only if the date column contains a date that falls on a Wednesday.

      1. Brewster's Angle Grinder Silver badge

        C'mon we're devs - not users

        Isn't that what your choice of scripting language is for? Ideally with somebody else's CSV parser or one you've written yourself and battle hardened over decades.

        This has the twin bonus that the script won't miss an entry while eyeballing a million rows and you have proven code ready to handle the file the next time it turns up. I've just spent the morning working on a 25,000 row TSV file. I could have hand hacked it, but the transforms I've made are documented in the code and if the source updates (or we want to use a bigger source file) I can run the transform again at zero cost.

  8. George Spiggott
    Facepalm

    Why anyone forks out for Office 365 when the excellent LO can be had for free I find bemusing.

    1. Fred Goldstein

      Compatibility. Everybody exchanges Word docs. LO comes close but risks breaking non-trivial things.

      And for some users, Access. It's old and limited, but what it does it does well. LO has no usable database program, just a bad joke called Base which chokes on anything more than a teetotaler's wine list.

      But Calc is pretty good, and Excel 2019/365 just sucks wind big time, buggier than hell. I am starting to convert things over to it just because it works and Excel often doesn't any more.

      1. ovation1357

        Sadly I have to agree although I blame Microsoft for the compatibility problems owing to their hideous non-open 'open' file format and their combined refusal to take ODF standards remotely seriously.

        I use LibreOffice 99%+ of the time now but just occasionally I have to fire up Office 2010 to open a document or spreadsheet which LO can't quite handle - this is usually either because of wanky formatting and/or Visual Basic Macros.

        However on the subject of Access I'd say that no matter how 'well' it works - it really isn't geared up to sharing a 'database' beyond its creator and I personally think it should be shot in the head because of the fact that it's just easy enough to use that people go creating 'useful' databases which quickly turn from being a pet project into something critical data in it - scale that out to a whole load of like-minded end users all creating their own little pet databases and you end up with a big problem of lots of vital information being holed-up in undocumented, uncontrolled and potentially un-backed-up, fragmented databases across an organisation many of which will have become bespoke applications with custom forms and a Mickey mouse front end UI..

        Nope, I say let's kill Access with fire and encourage people to seek a proper solution to their problems.

    2. W.S.Gosset Silver badge
      Unhappy

      LO is a magnificent credit to the unpaid coders doing it in their spare time.

      But vs MSWord or --CHRIST!!-- MSExcel, it's like a Trabant dressed up with racing stripes.

      Imitative graphics do NOT create functionality.

      If you're doing non-trivial work, you run so far so fast off LO's capabilities, it makes your head spin.

      It's like putting a Tai Chi "black belt" into the Octagon. "Whooooo, look at ME, I am throwing all these SHAPES, look at me styling, I have all the MOVES."

      *PUNCH*

      *splat*

      1. ovation1357

        This is massively disingenuous. LO fares pretty well against both Word and Excel, possibly not so much against PowerPoint because Impress is definitely behind. But then having just had to fight with genuine PowerPoint corrupting files and crashing today, I'm not convinced it's really worth it for the flash animations.

        I've worked on some fairly hefty technical documents in both Word and 'Writer' and I had far more limitations and problems with Word - random formatting that would result in having to delete a whole section or page and recreate its contents, and crashes/corruption when dealing with heavily commented files under change tracking. I don't think Writer ever crashed on me and it gave me minimal grief.

        Unfortunately, docx file compatibly problems meant I had to use Word when everyone else was but I'd happily binned it for Writer in a flash.

        Excel is a different kettle of fish - I don't think anyone disputes that it's got more features than Calc but all of the core functionality is there in Calc - even pivot tables. There's nothing your average Excel user wouldn't be able to do just as well in Calc.

        For me, a massive bonus of LibreOffice is that I don't have to use the Ribbon interface. For me it is miles easier to find everything I need from the menu bar.

        If there was just one feature from MS Office that I'd like to see in LO it would be the Font selector from Word which not only shows an example of the Font but will temporarily change your highlighted text to the font you're hovering over as a kind of live preview.

        1. W.S.Gosset Silver badge

          2 very different aspects here so I'll split my reply in 2. Part the First: (to paraphrase Douglas Adams)

          I agree with most of your objections to current MSOffice.

          Massive ++ re the ribbon interface! An absolute excrescence and a vile festering gangrenous fuckup. Hate it!!! Never met anyone who thinks otherwise.

          Fortunately the coders quietly kept all the menus under-the-hood so the old Alt keystrokes still work. Eg Alt I N D to get the Name window in MSExcel, wherein you spend a lot of time in any major modelling work.

          And that's a key point. Old versions.

          See, serious workers keep and use the old versions because they are so much more functional. It's like MS binned all the core architects after 2000, and everything since then has been a marketing-dumbarse directed charge backwards. Excel in particular peaked significantly at 98/2000. God the nonfunctional misery of 3D graphs now... And the catastrophic bandaid stickontop that is the "new" Tables "feature". Which is in the presentation layer (like 99% of everything added since 2000) so it's a deadend; can't build on/with it. And irrelevant tat if you understand Excel's intersections.

          Essentially, MS has just been vomiting graphics and instability onto MSOffice for 20yrs, while also actively REMOVING capabilities and functionalities.

          Tip re Word: yes, corruption in big files has always been an issue; you can sharply reduce your exposure by switching off the Fast Save option (on by default). If you open the file in a text editor, you'll note a Fast Save file is actually now a bundle of 2 files, the latter an appended transaction log reassembled on the fly.

        2. W.S.Gosset Silver badge

          Part the Second

          > massively disingenuous

          No. Quite the opposite.

          You just haven't tried to solve, or even have never encountered, the problems I and many others use these tools for. Weighted average portfolio risk on a correlation matrix? 1 formula in 1 cell, in MSExcel; in LO, a vomitous spray of 1 or more additional hardwired sheets plus location-/assumption-hardwired links farted throughout, with massive eye-watering bum-clenching fragility.

          The average tech-focussed individual quite literally has no idea how shallowly they're scratching the surface of the tools. For the simple and non-negative reason that their normal day involves solving very different (and to my mind much harder) problems.

          The reverse is true too. Try explaining to a quant or admin why vagrant is awesome for development. "Oh ok it's some sort of VM thing. So what." Or to an accountant why shell kicks arse. Then YOU blank out as they explain that unregulated transfer pricing is a major driver of globalisation rorting the smaller countries, if you understand how it's used.

          Key Point: Until you have a problem to solve, you don't fully understand the various tools' relative usefulness or value.

          E.g., I couldn't stand vi when I first had to use it. What a piece of shit! Then I had to do a lot of major wide and deep work in a hurry on raw Unix installs, HAD to learn its details to solve peculiar Deep problems, and promptly fell in love with it. Even today 30yrs later, I'll casually do something in 15secs and modern admins who'd scoffed at my offer to do their half-hour ballache for them in vi (and they laugh) will lean forward and say "how the hell did you do that so quickly".

          E.g., trying to explain to a DOS Command File guru why unix's shell makes it look agonisingly broken, brain-dead, and incapable of doing anything non-trivial, but them pointing out that it's feature-for-feature identical with shell.

          Unless you're trying to solve a particular problem AND have used the competing tools back-to-back for same, you won't really grok what that problem's problem-solvers value.

          But what are scientific writers doing using Word in the first place?? Wrong tool. Should be using LaTEX. Word's only ever been intended for smaller less complex stuff. LaTeX, FrameMaker, InDesign etc for the big stuff. Hourses for corses.

          1. W.S.Gosset Silver badge

            Re: Part the Second

            To get a glimpse of the sort of stuff that has Excel users ditching LO after 10mins of growing frustration and disappointment (kinda like going to walk through a door and bouncing off a photo of one), read my posts below this link below, which embed some low-level basics on which we daily casually rely:

            https://forums.theregister.com/forum/all/2020/06/30/libreoffice_70_beta/#c_4062568

            1. ovation1357

              Re: Part the Second

              Part the third (kind of)...

              Thanks for taking to time to put back polite, well reasoned replies - if only all other users could be so professional.

              You're absolutely right that I haven't encountered the kind of problems you're describing. When I've got some data to analyse I tend not to reach for a spreadsheet in the first part - depending on the problem I'm solving I might write a script, a bit of code, do some hacking about in vim or even import it into a temporary MySQL table where I can very quickly produce what I'm after. For me, spreadsheets are a last resort.

              I was doing my accounting in a spreadsheet but that was too clunky. Now I use an accounting package which is much better. I think in many cases people reach for a spreadsheet as it's generic and handy even though the task they're trying to do would be handled much better in a more specific application.

              Your example of =SUM(A:A * B:B) is quite an interesting one and I grant you that it both looks quite useful and also doesn't appear to work as-is in LibreOffice Calc right now. However my point is really that in any organisation the default is usually for all staff to have a copy of MS Office which comes at an enormous license cost and yet probably 90% of those users will never step beyond the current capabilities of LibreOffice, so why not save a small fortune on MS Licenses - encourage users to adapt to LO and make it the corporate default and then just buy a handful of MS Office licenses for the users who actually need its features?

              The ability to create an all-powerful super formula in a single cell is cool in one hand but sounds like it's also dangerously stepping into the territory of the intelligible Perl one-liner :-\

  9. ovation1357

    Large spreadsheets

    Not so long ago I encountered an error opening a spreadsheet in Calc that it had too many columns (I think!).

    Although in my case it turned out that the sheet contents didn't really extend beyond a much smaller count and it actually worked fine, that didn't stop me googling the error.

    It turns out that there is a fairly long-standing RFE against Calc about this limitation which they've been actively working on over a couple of years - I think this is related to the experimental large sheet support referenced in the article.

    IIRC as well as the column number being stored in too few bits to grown beyond something like 16k there was also a whole tonne of refactoring required to adopt a bigger number as well as rearchitecting the memory allocation such that it doesn't try to load the entire massive sheet into RAM all at once.

    It's really great that the LibreOffice team are still putting really significant effort into making it a better product.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2020