Google Bugs

This section was originally part of something like a blog. To contact me or comment on this, see my email page.

I started updating it again in early 2010. Further updates will be announced on my blog's latest updates.

  1. Google Search Bugs
  2. Adwords and Adsense
  3. Gmail/Googlemail, GoogleGroups and Orkut Bugs
  4. Google Desktop, Web Accelerator and Toolbar Bugs
  5. Google Docs and Spreadsheets
  6. Google Public DNS
  7. Google Software Project Hosting
  8. Blogger and Google's Blogs
  9. YouTube
  10. Google Business Methods Bugs
  11. Conclusion
  12. Other notes

photo


Adwords and Adsense

Adwords and adsense looks like some of Google's top mass-market money-makers. I have mixed reports about them, including some of my customers that are happy to use them, and others that get results like this commenter at Problogger which suggests that adsense referrals maybe don't always report accurately.

2007-07-23 (Permalink): Google seems to be having a harder time with adverts in 2007. Firstly, the Australian Competition and Consumer Commission alleges misleading and deceptive conduct by Trading Post and Google. Then, a report suggests Click Fraud Up to 15.8% in Second Quarter.

Now, it looks like Opera may be about to break Google adverts from many mobile devices - Harishankar asked:

"Does this blog post [Mobile Phones and Mobile Games: Google's difficulty in over-coming Opera mini's Proxy Server] saying "opera mini 4 is bringing difficuilties to google adsense" make sense?"

I'm surprised Opera's proxies are so privacy-friendly, but it makes sense to me. I expect Google will change their adsense policies yet again soon, if this worries their clients and they don't want a showdown with Opera.


Essentially, google is making the web more stupid, polluting it directly and indirectly with bad habits and bad business methods. The direct pollution comes from google's own services, such as gmail/googlemail base64-encoding far more than necessary, as described below.

Google's search is the core of their business and seems to be the least buggy service, causing the least direct pollution, but it is a source of indirect pollution. Web site owners try to "win" the latest reranking, because most users stop on page three of a search [BBC]. This wouldn't be so bad if the search results were halfway stable and not open to much manipulation.

But it is bad, because - even if you're not in China (see the China controversy down the page) - you can't rely on google's results from one minute to the next. The ranking algorithm is manipulated by Search Engine Optimisers (SEOs) as a paid service, and by google's admins as a punishment of sites they don't like. See this google guy's blog for one example of them ranking a site off by hand and then putting it back again [from the same blog] a few days later. A year later, it happens again: here's another consultant spotting lots of turbulance in the results. If Google's actions were unfair and harmed your performance, you can't touch them, as kinderstart discovered (but that ruling might be an interesting reference for other ratings sites, so there's a silver lining).

This frequent fiddling - and its dramatic effects - encourages the bad SEO sellers /spammers /scammers by letting them show the occasional "quick win" in the rankings, which helps them convince more unsuspecting victims to buy their pollution services. Google's probably not too unhappy with that, because most SEOs (good and bad) also buy Google ads for their services and clients: giving SEOs results for their adverts feeds the spiral and helps Google's profits soar [BBC].

The obvious ways to stamp out this pollution are:

  1. to start using bookmarks again, either directly or through shared-bookmark services - as well as reducing google's influence, this has the added benefit of helping reduce the typo-domain-squatters, and maybe help reduce "phishing" attacks (if you only access your subscribed sites through your bookmarks, you're less likely to end up on a fake);
  2. to help rebuild the directory services like vlib and dmoz - these sites are compiled by reviewers and sometimes give better results, but they always seem to need more help to keep up with the huge search index robots; and
  3. to use alternative search engines as a last resort.

We've got to stop searching and start navigating again, before we all drown in crap.

There is also a privacy/trust problem with having one company seeing so much of the world's searches. This was the basis for a recent Ask.com ad campaign (nice idea, shame about the microsite's poor accessibility and various other things - it gave one Google-worker a good laugh and got better commentary from Susan Hallam Communications). Privacy bodies backed Google's step towards allowing user privacy [BBC] but there's more that should be done.

Privacy International gave Google its worst rating, calling it "the most onerous privacy environment on the Internet" in June 2007 (seen at Lick the light switch). Matt Cutts's response seemed to look at what the other companies did wrong, rather than explaining why PI was wrong about Google. (The one example offered, Summer of Code, is a marketing campaign that also means they get new data about lots of students and mentors.)

Maybe reports of the PI rankings went in a bit hard on Google, maybe the other Big Search Engines are equally terrible and maybe there are more worrying things to investigate, but that doesn't mean PI are wrong about Google being a danger to your privacy. You simply can't see all the data Google holds about you or control what they do with some of it. When Google makes a small improvement in privacy, it is trumpeted from the rooftops, but why isn't it in there from the start?

Google Searches for English English

Patrick Alstrom highlights how Google was trying to change the English language even in the UK version of their site. As usual for the search site, it changes from time to time, so this may not happen when you read this.

Two small points that often come up about this:

  1. This bug isn't unique to Google (it stopped me using Ask.com when they made auto-correct hard to avoid).
  2. I disagree with the OED about the old-fashioned z spellings being more correct.

Gmail/Googlemail, GoogleGroups and Orkut Bugs

Due to privacy considerations and more, I will not respond to gmail or googlemail "private" emails. Please resend your email from a different place. For more detail about this, please visit Google Watch.

If you didn't know gmail is also called googlemail, then maybe you missed the reports that google ignored the gmail name already used in the UK and Germany [BBC]. I guess we'll see whether Google can burn a financial data firm in court...

Googlemail and Orkut use invite spams. Google seems to think that I should register each and every one of my addresses with them to avoid the invite spams. That's dumb. I have an infinite number of addresses, thanks to debian and others.

The invite spam problem may look like it's the fault of users, but so many are doing it! Google's advice about invites is totally inadequate and I think users who follow it may end up breaking privacy laws in some countries. I think Google probably wants to encourage invite-spamming: how many people will report their friends and colleagues to local privacy law enforcers? Classic externalisation of advertising "pollution" costs.

Despite its own spam creation, Google is keen to be seen as helping to fight spam. To that end, it files invalid spam reports like this report [Erich Schubert]. As Erich writes "That totally sucks, that a perfectly valid and correctly sent email, that is also delivered correctly, is reported as spam amongst the big ISPs."

There's little reason to use gmail/googlemail. You definitely shouldn't use it as your primary email account, as it's an example of unreliable services described by Ian Murdock in Dipshits like me and What does it mean to be a "paying customer" in Web 2.0?.

Googlegroups changed its interface in 2005, causing a whole new load of problems for Usenet, almost as bad as when AOL connected (but AOL's now left again). Replying In Google Groups [Safalra's Website] explains the worst change. Others include requiring registration to see email addresses. It shouldn't be a surprise. Even when it first launched, there were Bugs in Google's Newsreading Service [RJK] which are still unfixed and unanswered, by the looks of it.

Gmail/googlemail is not fit for mailing lists either. Did you know that gmail has a bug which makes it base64-encode text emails for little reason, making your message about 30% larger and unreadable to some users? This bug has been reported to google, but they haven't yet even acknowledged it. I strongly suggest using a different webmail - one that fixes bugs. Testing seems to suggest it does it for non-ISO-8859-1 (West European) users more than anyone else.

Security bugs seem to take a long time to be fixed, too. There's a running sequence about this in the Jibbering Musings, including Google Flaw not fixed, GMail contact stealing demo.

Maybe they don't acknowledge my bugs because some google mailservers reject entire TLDs - it seems that I can't use my .coop addresses for googlegroups, for example. What does google have against cooperatives?

We may never know, as there is no list of known bugs. If you guess the correct keywords and email them in, maybe you can get a canned reply confirmation like this blogger [Life and Microsoft].


Google Desktop, Web Accelerator and Toolbar Bugs

Update: Featured on c't magazin TV show week beginning 6 May.

There's a more obvious privacy problem in Google Desktop on Microsoft Windows. Unless you stop it, it will send your hard drive contents to Google's servers [delete the border]. This is the sort of thing I would like to see made "Massive Computer Misuse" in any updates to English law (as mentioned in my submission to the All-party Parliamentary Internet Group DRM enquiry). I'm sure it was in subsection 275 of the terms and conditions, but I'm also pretty sure that it was presented so that most of its victims didn't see or understand it.

Apparently, the Google Web Accelerator [is] considered overzealous [ORA] and it can combine with Javascript-abusing web applications to cause data loss. I think someone could have forseen what could happen when you start prefetching all links, but it happened all the same. So, what do you think the consequences of sending all your data to google might be and will it happen?


Google Docs and Spreadsheets

January 2007: A relatively recent addition to the Google range is Google Docs. It has already suffered google's usual security/privacy bugs and is at risk from similar problems as the other hosted services, like Google Mail.

June 2007: I borrowed someone's web browser and tried Google Docs and Spreadsheets, thanks to an invitation from Will Pollard. I borrowed a browser because it simply doesn't work on Any Browser and left me in a page with no working links. I hope that gets fixed before it leaves beta. You can't even view a document you've already created.

Using a full-fat browser, it seems to work, although I could hear the processor fan spin faster under the extra load. It's probably not as processor-intensive as running OpenOffice yourself, but I think it's comparable to a lighter word processor. Forget about web applications moving all the processing to the server and saving you from the upgrade treadmill: this doesn't seem a lightweight option to me.

The Docs editor interface is a page with a Google header, then a rather confused bar, then a toolbar with links to drive a TinyMCE-style editor in a textarea below. At the bottom is a status bar. The confused bar contains dropdowns, tabs and links with little reason for which is which, as far as I could see. In fact, some change from on to another when actived. One of the link/tabs is an 'Edit HTML' that brings the document source into a text box - non-Javascript browsers could be shown that instead of a near-blank page, instead of being left stranded as I mentioned earlier.

Another thing that seems to be confused still is the software's name. It calls itself Writely at some points, which was the company Google bought for this, not long after the were cosying up to Ask.

The site looks like it won't work on devices with small screens. If the browser is too small horizontally, some controls are simply unreachable. The status messages for saving and so on don't appear in the status bar, but over the Home/Logout links in the top-right of the header. This seems another aspect of the interface that still seems very beta, very confused.

However, within its limits (full-fat browser, big screen, fast computer, fast connection), it seems to work OK. I couldn't trip it up with utf-8 and a few other nasties, but sometimes the editor is slow to respond and I had the whole publishing page grey out on me once.

Once published, the page is visible with any browser, but it is not a valid web page - my simple test page failed the W3C validator with 36 errors - impressive for a two-line page. The exports as Text, PDF and OpenDocument seemed OK, but not much better at first glance than Mozilla or OpenOffice will do if you give them the html, and you probably have those already if you can use Google Docs and its OpenDocument export.

In conclusion, I don't see the attraction or the reason for the hype. If you want to post web pages, using your own editor or putting a wiki on your web space or getting a free wiki is much easier to use, easier to bugfix and might even produce valid web pages. If you want to use a Word Processor, using one on your local PC is not much more processor-intensive and things like OpenOffice will probably produce better output.


Google Public DNS

January 2010: Another new service. Rick Moen in LG #170 discusses Google Public DNS in some detail. So far, there seem to be only concerns, not actual bugs. The software appears to have been written by a DNS programming novice, the motivation isn't clear, and users won't have the contractual protections they have with most ISPs.

Alternative: Rick Moen advocates running your own DNS if you can. Otherwise, let your router and computers use your ISP's recommendation automatically, like they would usually do. You'd need to make an active choice to get weaker DNS service from Google, so why bother?


Google Software Project Hosting

Google's hosting service seems to share Sourceforge's non-free code problem. I've not many reports about the hosting yet, but I'll be surprised if old problems like the registration pressure, accessibility failures, .coop bans and so on from above aren't repeated.

I'd avoid Savannah and GNA for now, unless you know you are happy with their policies (such as required early adoption of FDL, or HTTPS-only). BerliOS.de and TuxFamily.org seem more relaxed, but that does mean you have to check each licence - a problem for browsers more than publishers, I guess. If there's a debian angle, alioth.debian.org is another option, or eduforge for education-related things (thanks to Penny Leach for the eduforge reminder).


Blogger and Google's Blogs

Apart from encouraging the larger Atom feed format, dropping RDF support and unreliable alternatives to its eyetests, Blogger is one of Google's best services. The basic service seems to work in a half-decent manner. Of course, it still has bugs and that let someone post a hoax Google news release

Update 2010-02-10: Google has bought reCAPTCHA and is pushing it to webmasters even though their own blog service does not use it yet. Don't be conned: reCAPTCHA is disability discrimination, not an anti-spam test. reCAPTCHA does not examine the form contents for spam characteristics at all. There are much better options out there, including the blogspam cooperatively-developed service.


YouTube

How annoying is YouTube? The site complains constantly that I don't let it rampage through my browser and I haven't downloaded the latest from Adobe. It's a bit less annoying now that ytplay is available, but it shows that Google doesn't care much for free software support.

An interesting twist has been Viacom suing Google for a billion dollars and calling them all sorts of names. (Beware bias in the BBC reporting on this, as BBC just signed up for Google-YouTube [BBC] - is there a monopoly that the BBC won't befriend?)


Google Business Methods Bugs

Some google-fans have started using google as a verb, as in "to google" instead of "to web-search". I've not been doing that and now it seems Google is sending legal letters asking people to stop using it as a verb [The Independent]. Google wants to be part of the language, but not too much.

Worse but less obvious, Google is a sponsor of the World Economic Forum [WEF site], where corporations tell governments how to control their electors to benefit businesses. That's the wrong way round: corporations should be subject to democracy and accountability. If you've not noticed the harm WEF is doing around you, look on social media and some of the non-violent anti-globalisation sites. For more information, see Joel Bakan's book "The Corporation" and this VIRUS article.

Google are using the old "give us your personal info to use some services" tricks. It is possible to use most of them without registration, but it's made more difficult than with registration. It's all part of what google fans correctly call "total worl domination" [tribulaciones]. They're not alone in doing this, but "User lock-in is not a business model; it is a mentally ill infestation that shall be eradicated" (from Closed for Business: google, MSN, Yahoo [Funkyware: ITCetera])

Google uses software patent protectionism so even the bits out in the open are kept closed, such as patenting searching with a voice interface [Pocket-lint]. I have another page with recent software patent news that includes links to explanations of how software patents and other "New Enclosures" can harm free software.

Google's Spin Attack

Google spins like crazy to deflect criticism of its business methods. It formed the Google Foundation, putting it in illustrious company with Nestle, Shell, the Gateses and a long list of others using its soaring profits [BBC] to buy goodwill from non-profits.

To win over sceptical hackers (skilled programmers [FOLDOC]), it has released some of its less profitable software as free software on Google Code.

It also runs a "Summer of Code" scheme which buys some projects a student worker for the summer. My view is that we should note they "aren't a particularly good example, but if you can get money out of them to improve debian and the web search tools it contains, good luck!" (from a debian-project email).

Google and the China Controversy

3 Feb 2010: spotted Google in China: Unanswered Questions, by Danny O'Brien at EFF.

15 Jan 2010: Google announces that they "will be discussing with the Chinese government the basis on which we could operate" (thanks Dave). This was reported on the mass media as Google 'may end China operations' which is possible but seems unlikely. After all, what has really changed in the general situation?

25 Jan 2006: google is taking a lot of heat online for agreeing to Chinese government content requests. I won't criticise the actual decision too much, because it's typical corporation behaviour [film site]: follow the money, like the rest of the World Economic Forum. If you're buying Chinese products just because they're cheaper, you're part of the reason they have the money and part of the reason that google is following them - corporations are seldom held accountable. If you don't like that, maybe you should Boycott Made In China as well as google?

As noted in TunaSpecial's Google Redux, MSN also complies with censorship requests and Bill Gates defended Google when the question came up at WEF. Once this stink also started to stick to the other China-censored search engines, they tried the face-saving move of asking the US Government to help [bink] - like that's going to happen while US shoppers prefer cheap Chinese consumables. Just remember the flak EU Trade Commissioner Mandelson took for enforcing the anti-dumping rules.


Conclusion

If you believe that privacy matters, that bugs should be fixed (or at least answered) or that WEF is wrong, say no to google. Boycotting it will help to make the point.

Students for a Free Tibet are keeping a list of google alternatives and Alternative, innovative search engines [madduck's droppings] is another list. Cutting Free reviewed ippimail: Free Software web-mail with a feel-good factor as a googlemail replacement.

There are many uses for people search engines, obvious and not. Many colleges and universities have an internal people search engine to find alumni and staff, but you'll have to use other people search engines to find old classmates or neighbors again. +

Steev commented:

"Nothing against ippimail, I do like the fact that it uses ad revenue to fund charities, even of my choosing, however, the difference between ippimail and googlemail IS the interface. No offense to the developers, but squirrelmail/horde really needs to be updated. There are many issues with it, which is why any time I use a webmail service that has squirrelmail/horde installed, I immediately look for a way to download the mail and read it in a mailclient of my choice, the thing is - that is going against the whole point of webmail. I haven't researched a lot of your arguments against Google, and I plan to, but just wanted you to know that even though I like what ippimail is doing, I simply cannot stand the interface."

I can't comment much on that: I've used squirrelmail without problem in the past. I tried to look at googlemail soon after its launch and was defeated by its totally-WCAG-busting interface.

Simon Martin responded with a comment:

"Steev, The interface to Squirrelmail is being updated slowly. To speed the process, help spread the word about ippimail so we can afford to spend more time/money on it."

"Hula also seems to be gathering momentum since the community got their hands on it properly..."

Tom Chance notes that the Green Party in England and Wales has motions put to its Autumn 2006 conference to boycott Microsoft, Google (both for supporting the Chinese info-goolag) and Yahoo (supply of evidence to China and censorship of trade union adverts).


World Economic Forum coverage

In 2006 as in past years, satellite TV channel SFinfo at 13e will be broadcasting some sessions of the World Economic Forum, with both German and Original (often English) soundtracks (I think Left/Right split). If you have a satellite set, you can see what the barons of big business are saying: sometimes depressingly hilariously irresponsible.

At 16:00 UTC tomorrow (Friday 27 Jan), the session WEF Digital 2.0: Powering a Creative Economy is televised. Now, the only one of that panel who doesn't alarm me is the first one and that's probably only because I don't know who he is. Gather round, one and all, get ready to chuck the rotten fruit at the TV.


Other notes

Google Blogoscoped analyses Google's latest moves in amazing detail.

Amusingly, a Yahoo! search for google problems returned this google answer about a urination problem high up in the list (fifth result or so). Now that really is taking the proverbial...

This is copyright 2006-2007 MJ Ray. See fuller notice on front page.