This archive is retained to ensure existing URLs remain functional. It will not contain any emails sent to this mailing list after July 1, 2024. For all messages, including those sent before and after this date, please visit the new location of the archive at https://mailman.ripe.net/archives/list/[email protected]/
[members-discuss] GM topic
- Previous message (by thread): [members-discuss] GM topic
- Next message (by thread): [members-discuss] GM topic
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
ivaylo
ivaylo at bglans.net
Mon Apr 29 22:57:09 CEST 2024
Thank you very much for your opinion Doru Serdin ! Thank you very much for your opinion Kaj Niemi ! Thank you very much for your opinion Kai Siering ! Down here wrote comments to every one of you (combined in one mail to not bloat the list). > "1. If you have 3 own hungry childrens and just one apple how will you > distribute the apple to them ? " > Your analogy is ..... lacking. > RIPE members are not hungry children who can't get their own food and RIPE > is not anyone's caretaker. > Resources are held by RIPE members (AKE the children) and transfers are made > every day, some people just don't want to pay for IPv4 simply because in the > past they were given out for free when needed. Some members not are only hungry, they are starving for INR (waiting queue list). My analogue is as much as close to the current situation, if you have better one explain it please. RIPE NCC always were, are, and will be care taker / arbiter (in case you close your LIR, what will you do with your ASN and IPV6 ? will "sell" them too ? or will return them to RIPE pool for redistribution ?) > You appear to simply want IPv4, not want to pay for said IPv4 and are > looking for changes that would benefit you at the expense of those that > already have that IPv4 available. Do you have any proofs for these your words ? I appeal you show such proofs publicly here ! Otherwise we can assume your statement as slander and you as lier. Luckily this mailing list is public and all messages are logged, you can go find and show us with quote, which I have wrote that these are my intentions. > Now IPv4 has been distributed already, it's gone, deal with it. Either use > IPv6 or buy/rent IPv4 from other LIRs. The solution exists for everyone who > wants IPv4, it's called money. No the [re]distribution process is endless process. For all INR there will be always in/out actions from the pool. From a registry perspective, it makes no difference whether we are talking about IPV4/IPV6/ASN. All INR should be treated by the same universal rules. If we have a shortage of IPV4 today, it could be also ASN or IPV6 tomorrow. > I on the other hand only want RIPE to spend less money in order for everyone > to have to pay less in membership fees in the future. > If you want to have a membership fee based on held resources, then the > voting power for each LIR should be proportional to that membership fee. I have a proposal for you: Transfer all your LIR resources for free to my company LIR. We will do a contract with your company, and I will provide you registry services for *some* time. We will charge you with 1/4 from our LIR anual fee and will give you 1/4 from our vote right. All aspects of your logic will be filled: 1. Company to Company contract for INR, and not RIR to LIR (LIR to LIR contracts about INR for money) 2. Less fee for your company. 3. Propotioning voting power of your fee. Do you agree mr Doru Serdin ? ---------- To: Kaj Niemi > RIPE-639 specifically mentions in its scope that rights to hold, use or > transfer (and remember transfers do not need to be free) "are not > addressed or restricted" by it. It also mentions that any existing or > future policies do not apply to legacy holders unless it is explicitly > included in the other policy. Citate in point 1.2 from ripe-639: "Any existing or future RIPE policy referring to resources shall not apply to legacy resources unless the policy explicitly includes legacy resources in its scope." And because I am talking for a change in ripe-639 (update) my statements are prety valid. ----> Change in policy that include legacy resources in its scope. Another one point here, in the next 2-5 years we the network operators access/content ISP, maybe will completely drop to rely on an old inetnum/inet6num routing information and will switch entirely to ROA/rpki mechanisms. So what will be the purpouse and usage of these legacy resources then ? No one will forward to/from these networks, so why should the register still support them ? The purpose of RIR registers is to provide adequate information about GLOBAL INTERNET PUBLIC RESOURCES wich operators to use for guarantee normal internet work and fraud prevention, not for registering someone's private intranets. > RIPE-822 mentions that assignments are valid as long as the original > criteria the assignment was based on is valid and the assignment is > registered in the database. It also mentions something about fairness. You are refering to point 6.3 wich describe LIR to end user assignment. We are talking for point 5. RIPE NCC to LIR allocations. > In practice, whether you want it or not, holder equals owner and holding > equals ownership. All laws and lawers in EU and in 99% of the world will tell you right oposite. You dont own part of the road on the highway where your car is, just because you pay tax your car to be on the road. You can use it, hold it (for some time), but you can not go drill it dig it take a part of it and sell it, because your car is on top of it and it is "your" property. > One cannot sell or rent something that one do not own or > have the rights to. Doing so would constitute some kind of fraud. People > sell and buy and rent these daily. In practice it is kind of a fraud indeed. But it is regulated by contract between companies / personals. If I want somebody to pay me for the air he/she breath, finding person whom agree to pay me, and then we make official contract, how to classify it ? Human stupidity is limitless, but to what extent exploiting it is legal or legally prohibited is a stretch question. > As for the apples. 1. in three or 4 parts depending on whether you're > willing to sacrifice yourself in favor of your progeny, 2. you should look > into buying futures to lock in the price of apples as you are producing a > lot daily (not financial advice), 3. you buy apples from someone else. For 1, now you know what RIPE NCC must do with the IPV4 space, and as a parrent you will cut the apple on 3 equal pieces, because every normal parrent love equal all its own kids, and always is ready to sacrifice self in favor of progeny. For 2 Even it not very standart answer, it shows what have to be made with IPV6 and ASN INR, RIPE NCC should lock part of his budget and on that resources too. for the future when IPV4 will be totaly unusable. Also soon or later you will go to question 3 (nothing is forever, soon or later your tree will die - IPV6/ASN will be exhausted) For 3 You can not just buy apples from somewhere, INR are not magicaly show up from somewhere and somebody to sell them to you, they are finite numbers. To produce new INR you need decades in developing, clear and implementing new standarts all over the world (as you can see with IPV6 integration), and another decades to make everybody start using it. Thats why we also need a guard mechanisms to prevent wastage and rapid unreasonable depletion. Now combine all and find common point, with wich we can achive all key targets. And tell what strategy we should have to complete the goals. ---------- To: Kai Siering All facts you are showing are true, but such a policy was a mistake, and we _must_ not repeat the same mistakes again and again. Today the main focus is on IPV4, but tomorow these talks will be applied on IPV6. Then again LIR-X will say... "Hey why LIR-Z hold /4 IPV6, and pay same fee as LIR-X wich holds only /32 IPV6 ?". In one of previous mails someone gave example with cogent 200M flow from INR (btw such sum is pocket money for one of the biggest T1 carriers, and such huge company), and from 200M do you beleave they can not contribute 1M to the register from wich they earn ? We need clear fair INR distribution rules / policies / charing fees, not related of the time line or resource shortage / previous historical distributions. We need RIPE NCC to be resource distribution arbiter between us, because human nature is to want more and more, to be greedy. RIPE NCC have to start look at its LIRs as to own childrens, whom can not be separated on categories or be disriminated on some signs. This is the only way RIPE NCC to have stable future and guarantee propper budget to operate. Otherwise the RIPE will lose the trust of his own members. That is what I am fighting for, loosing my time, and writing tons of emails here. Ivaylo Josifov VarnaIX / Varteh LTD +359 52 969393 Varna, Bulgaria On Mon, 29 Apr 2024, Doru Serdin wrote: > "1. If you have 3 own hungry childrens and just one apple how will you > distribute the apple to them ? " > Your analogy is ..... lacking. > RIPE members are not hungry children who can't get their own food and RIPE > is not anyone's caretaker. > Resources are held by RIPE members (AKE the children) and transfers are made > every day, some people just don't want to pay for IPv4 simply because in the > past they were given out for free when needed. > You appear to simply want IPv4, not want to pay for said IPv4 and are > looking for changes that would benefit you at the expense of those that > already have that IPv4 available. > Now IPv4 has been distributed already, it's gone, deal with it. Either use > IPv6 or buy/rent IPv4 from other LIRs. The solution exists for everyone who > wants IPv4, it's called money. > I on the other hand only want RIPE to spend less money in order for everyone > to have to pay less in membership fees in the future. > If you want to have a membership fee based on held resources, then the > voting power for each LIR should be proportional to that membership fee. > -- > > ____________________________________________________________________________ > Doru Serdin > Network Manager > Office: +4 031 82 52 657 > E-mail: doru.serdin at mediasat.ro > > www.mediasat.ro > > www.alonia.ro > > Privileged/Confidential Information may be contained in this message. If you > are not the addressee indicated in this message (or responsible for delivery > of the message to such person), you may not copy or deliver this message to > anyone. In such case, you should destroy this message and kindly notify the > sender by reply email. Please advise immediately if you or your employer > does not consent to Internet email for messages of this kind. Opinions, > conclusions and other information in this message that do not relate to the > official business of my firm shall be understood as neither given nor > endorsed by it. > > ? > > On 29.04.2024 3:18 PM, ivaylo wrote: > > I think it?s great we are being creative with the > ideation of the models but? > > Fair is a very vague yet complicated concept. Here I > agree with Nick ;) > > Distribution in the past has been needs based (based > on existing or > projected need). It's done. > > > I dont see how the word "fair" can be vague. When you dont have > shortage of the resource it purely means "depend on your needs" > (current situation with IPV6, ASN). But when you run on the > limit it means equal (IPV4). > > Let me ask you 3 questions: > > 1. If you have 3 own hungry childrens and just one apple how > will you distribute the apple to them ? > 2. If you have 3 own hungry childrens and 1000 apples, but > tomorow your apple tree can give you another 1000 apples ? > 3. If you have 3 own hungry childrens 1000 apples, but you old > apple tree is dead and that food is all you have for undefined > period of time until your new not seeded yet apple tree grow and > start giving you apples ? > > When you answer me of these 3 questions you will see and what > _MUST_ mean of the word -->fair<-- is for RIPE. And when you > find the common point of your answers you will know what RIPE > NCC must do. > > Kaj, you talk so much about lawsuits, but can you give a citate, > or point to RIPE policy or agreement, where it is said that > allocated from RIPE *INR to the LIRs are not subject of change, > not subject of their quantity change, and they are distributed > for undefined period of time ? > > If there are still people or companies which mistake the word > "hold" with the word "own" is entirely their problem. Nor RIR, > LIR or end user can "own" *INR. They just can hold and operate > with given INR for defined period of time (look at the RIPE > documents and contracts, then tell me where is used the word > "own"). The funny part is that the only way personal or company > to be identified as authority of the given resource is the > register itself. If the register stop functioning no one will > know which *INR you can operate with. So if you screw (lawsuits) > the register you are screwing and your own interests. > > *INR - Internet Number Resources. Any Internet identifiers such > as IP addresses (IPv4, IPv6) and Autonomous System Numbers. > (Citate from RIPE documents) > > One _OBLIGATION_ of RIPE NCC is to care for *fair* distribution > of the resources across the LIRs in the region. And am not > saying it is subject of member's wish or vote how the INR have > been / are / will be distributed, I am saying it _MUST_ be > fundamental rule that RIPE NCC must apply to guarantee its > future, and it must not depend on any contractor or requestor > wish. > > ---- > About RIPE-639 > > The policy: > https://www.ripe.net/publications/docs/ripe-639/ > > Mr. Xavier Le Bris report from 12.01.2024 > (Big thanks for his work and efforts): > https://labs.ripe.net/author/xavier/10-years-of-legacy-policy/ > > I dont see anywhere the RIPE register is _obligated_ to keep > legacy resources records for undefined time. I only read the > RIPE register is _authoritive_ for these INR. With other words > the data for legacy resources are holded in the register by good > will. And there is no any obstacle to _NOT_ have deadline to > keep doing this (in point 2.6). Also from the report 9% of 12 x > /8 IPV4 blocks (nearly full /8 IPV4 block) is with undefined > status. Another legacy holders which holds nearly 30% of the INR > are out of any contracts or touch. So where is the benefit to > keep holding such records ? Will the register be more acurate - > NO ! Will the memebers will benefit of unusing nearly /8 - NO ! > Are these legacy holders contribute something to the RIPE NCC - > NO ! Correct me if I am wrong please. > > ------ > About the claim "Last year base of the members rejected category > based charging scheme, we will not offer it again". As I > remember offered model B for 2024 also was not supported > (rejected) , why it is offer for 2025 again ? When we take a > decision what to offer and what not is rely on principle or on > personal subjective opinion ? > > For 2024: > https://www.ripe.net/media/documents/Option_B_-_Price_Increase_10_-_RIPE_NC > C_Charging_Scheme_2024.pdf > > For 2025: > https://www.ripe.net/media/documents/Option_B_RIPE_NCC_Charging_Scheme_2025 > .pdf > > To note that I and organisations I present are against category > based charging schemes which will put the main load over the > small to medium resource holders (LIRs), and will favoritize the > biggest (what was offered in the last year). > > > > Ivaylo Josifov > VarnaIX / Varteh LTD > +359 52 969393 > Varna, Bulgaria > > > On Fri, 26 Apr 2024, Kaj Niemi wrote: > > I think it?s great we are being creative with the > ideation of the models but? > > Fair is a very vague yet complicated concept. Here I > agree with Nick ;) > > Distribution in the past has been needs based (based > on existing or > projected need). It's done. > > Any real implementation of forcing a "fair > re-redistribution" of RIR > assigned resources, as you suggest, will lead in > lawsuits. Whether it would > be 1, 10 or 1000 doesn't matter. The association > will not have the financial > resources and time to handle those. Probably not the > best use of LIR fees, > either. I doubt anyone in legal would ever ACK this. > > Any of the companies holding address space that > RIPE-639 refers to might not > have an interest, or even be aware of the "benefits" > of being a RIPE LIR > themselves. Yes, they don't know about RPKI either, > usually. Legacy reverse > DNS might be enough. Many of these are > multinationals in non-tech sectors > who have gotten addresses before the RIR system > existed. Some of these got > addresses around the time classless routing came > along from their vendors > who themselves had gotten the addresses from IANA > just by asking. Some of > these do not route the addresses on the internet, > some others do, etc. etc. > Again, ruining someone's day by "reclaiming" - let's > call it what it really > is, hijacking - address space that hasn't been RIR > assigned will certainly > lead to costly lawsuits and claims of damages as > above. > > Don't think a "Pigouvian tax" would work either as > the alternative is always > for the holder to sell or rent the addresses. I > think it'll lead to more > consolidation of addresses to the cloud providers. > Probably again not what > was intended originally. Any added cost would > eventually be shifted to the > customers one way or another. Apropos that, Cogent > is raising a 200M note > secured by a bunch of IPv4 addresses and rental cash > flows. Addresses have a > significant value as collateral just as stable cash > flows do. > > > :) > > > > Kaj > > Sent from my iPad > > ___________________________________________________________________________ > _ > From: members-discuss > <members-discuss-bounces at ripe.net> on behalf of > ivaylo > <ivaylo at bglans.net> > Sent: Friday, April 26, 2024 9:26 PM > To: Massimiliano Stucchi <max at stucchi.ch> > Cc: members-discuss at ripe.net > <members-discuss at ripe.net> > Subject: Re: [members-discuss] GM topic > > Thank you for this example Massimiliano ! > > It is very good example how bad the resources can be > used, and how bad is > to have flat fee for all members ! That's why RIPE > fee _MUST_ include per > resource component. > > I am currious about technical aspects in the logic, > Even if you go > with EUI-64 standart for your network infrastructure > (nobody do this, > because the addresses unprediction) a /29 IPV6 will > let you have 34 358 > 689 800 ( > 34 Bilions segments). With the maximum > ttl value of 255 > (limited by the one byte in the header, current > around the internet is > used just 64), You can have 134 739 960 (> 134 > millions) router > interfaces. With the modern switching asics, you > will need and around > 144 000 000 000 000 (> 144 trillions) ethernet > switches where you can > connect up to 18446744073709551616 end hosts per > segment. Let me also > remind that IPV6 protocol allow to use a single > (just one) adress per a > system/ruter/host (not per interface as IPV4) > implemented in almost all > vendors equipment and OS. > > Yes in theory it is posible to need huge address > space, but in theory > every one of us (LIRs) have to drawn all ipv6 + ipv4 > + ASN from IANA (not > from RIPE only) to cover all very unlikely future > needs. > > ----- > > The problems RIPE NCC and we members have, is not > the charging scheme is > fair or unfair, the root of the problem is _UNFAIR > RESROURCES > DISTRIBUTION_ during the years. As everything in the > real world internet > resources are also finite numbers. So how to spread > finite resources to > group of members which by presumptions are equal > (First and most > important RIPE obligation - threat all members > equally) ? Logically - > equal. If we were 5 members maybe we can handshake > each other and have an > agreement (not sure ether when we talk about > concurent companies / > personals), but we are 21k that's why we need clear, > transparent and > strong rules equal to all. We currently have a > strong discrimination based > on very vague foundations and rules. > > I will be very happy somebody from RIPE to explain > us, why one member can > hold /8 just for fun (or testing or whatever), in > the same time other > member to have single /24 (or to be in a waiting > queues) extreamly > dificult to operate its bussiness because lack of > resources and both of > them to be forced to pay same fee. Which rule in the > RIPE founding > agreement exactly says this is fair and equal ? Also > how many from the > RIPE NCC staff and you LIRs, with hand on heart will > tell the current > situation is moral, normal and fair to all ? > > In all cases after the posible delegated resources > to RIPE have limits we > also need limits for the resource each of us can > hold and operate with. > thing more fair can be than to take all resources in > each category > (IPV4/IPV6/ASN) which are delegated to RIPENCC and > divide by the LIR > members number. > > > Solution one: > Redistribute resources equaly to all. Not imposible > as I already wrote, > but will that lead to tremors in the work of > operators - very likely. But > what is that is equal fees, equal rights, equal > resources. > > Solution two: > Start to "penalise" the the LIRs which holds and > operate resources over > the fair share limits. That will not make disruption > of the internet work > in the region. And will give time to those who want > to optimise the > resource usage to do it and save expenses. How to do > it ? - With money > ofcourse. To prevent adding too much initial stress > to all, we have to > start with small steps and calibrate the > "penalisation" fee with time. > > As I wrote in my previous mail the current fair > shair resources for each > LIR member based on what I checked (by IANA > documents and 21500 LIRs) are: > > 1 x /18 IPV4 block > 1 x /28 IPV6 block > 16 x ASN > > Everything above these numbers should be "penalise" > with a small fee per > each resource over the limit. Ofcourse these limits > must be shifted if > we become 40k members or down to 10k, but it is an > easy part. Also > "penalty" fee by 10 euro per year for /24, /32, ASN > looks prety resonable > in the light of current free market prices - 15k > "buy" / 100 euro month > rent. > > It is not a category based charging scheme ! > It is fair share charging scheme ! > Can be looked at like extension of the scheme C, but > guarantee much more > RIPE NCC sustainable future and covers all LIR > resources (not only PI). > Current offered option C is inanity, because one LIR > can hold /8 PA and > none PI space, the other can hold single /24 PA and > 10 x /24 PI, and will > pay much more than the first one with /8 PA (fair ?) > ! > > > Another important theme is RIPE-639. 10 Years after > the adoption we still > have nearly 34% undiscovered resources. How much > more time it needs ? If > for 10 years you cant find legacy holders what will > be the chance to do > it in the next 10 or 100 years ? We need a straight > deadline to point > 2.6. I think RIPE NCC can safely suspend (not > delete) all records for > these 34% resources. It is good to put dummy records > to point to call or > email to RIPE support staff, and these resource to > be gobaly announce. If > nobody call in few (3) mounths, can be consider free > and can be drawn into > the pool. > > > > Ivaylo Josifov > VarnaIX / Varteh LTD > +359 52 969393 > Varna, Bulgaria > > > On Fri, 26 Apr 2024, Massimiliano Stucchi wrote: > > > > > Hi, > > > > On 26.04.2024 16:40, Thibaud Perret wrote: > > > >> I have a question for Massimiliano, if he sees > it: > >> Do you really need that many IPv6 addresses? > Couldn't you just go with a > >> single > >> /40 allocation instead? That would still make a > couple of /48 to be > >> announced > >> separately for you to keep your "lab allocation". > If the answer is "yes" > to > >> that > >> question, you could then pay well less in most if > not all other RIR > >> regions. > > > > I could do with a much smaller allocation. At > present, I'm announcing both > > /29s, with the first one being where my > infrastructure is (and an address > I'm > > writing this e-mail from), and the second /29 has > been used for some > research > > such as the one you can find here > >(https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flabs.ri > p > e.net%2Fauthor%2Fstucchimax%2Fa-bgp-side-effect-of-rpki%2F&data=05%7C02%7C > %7C0b84b59374a24b82c60308dc661e588c%7Cd0b71c570f9b4acc923b81d0b26b55b3%7C0 > %7C0%7C638497527736209710%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQ > IjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C4000%7C%7C%7C&sdata=y2T5QV8Ea > 7AuXD4WBPSVxPxiuOcA80pYk7PGy8LHnyQ%3D&reserved=0). > > > > The point of the article was to compare what a > "normal" LIR would pay in > the > > different RIRs. If I were to compare having a /48 > (which would be PI at > that > > point) I would be comparing two different things. > > Additionally, I wanted to showcase how some > decision could be hindering > > Internet development, such as the one from LACNIC > to charge so much for > IPv6 > > compared to all the other RIRs, and the similar > situation at APNIC, > although > > to a lesser extent. > > > > To wrap this, could I do with a smaller > allocation? Definitely. But I > > received a /29 as part of my membership and I can > make some use of it. I > > could also live without a second /29, but that > does not increase my > > membership fees. I would return it if the charging > scheme were to be like > > APNIC or LACNIC. > > > > Ciao! > > > > -- > > Massimiliano Stucchi > > MS16801-RIPE > >https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstucchi. > c > h%2F&data=05%7C02%7C%7C0b84b59374a24b82c60308dc661e588c%7Cd0b71c570f9b4acc > 923b81d0b26b55b3%7C0%7C0%7C638497527736218822%7CUnknown%7CTWFpbGZsb3d8eyJW > IjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C4000%7C%7 > C%7C&sdata=CCE0El%2F5B41Uiky2BkscGB%2Fb%2FnkIWS2tKh4KGwPNA%2F4%3D&reserved > =0 > > > > _______________________________________________ > members-discuss mailing list > members-discuss at ripe.net > https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.rip > > e.net%2Fmailman%2Flistinfo%2Fmembers-discuss&data=05%7C02%7C%7C0b84b59374a > 24b82c60308dc661e588c%7Cd0b71c570f9b4acc923b81d0b26b55b3%7C0%7C0%7C6384975 > 27736224712%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLC > JBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C4000%7C%7C%7C&sdata=AKpL%2FBU%2BGivvG9mSvj2 > oFyeYeLXVit3cGEBi0iHep9Y%3D&reserved=0 > Unsubscribe:https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F > %2Flists.rip > e.net%2Fmailman%2Foptions%2Fmembers-discuss%2Fkajtzu%2540basen.net&data=05 > %7C02%7C%7C0b84b59374a24b82c60308dc661e588c%7Cd0b71c570f9b4acc923b81d0b26b > 55b3%7C0%7C0%7C638497527736228202%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAw > MDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C4000%7C%7C%7C&sdata=w > Gy%2BZKcmJxuRIwj9%2BTXb5a6zDlImgJ%2BgmCg51B4zlw8%3D&reserved=0 > > > > _______________________________________________ > members-discuss mailing list > members-discuss at ripe.net > https://mailman.ripe.net/ > Unsubscribe:https://lists.ripe.net/mailman/options/members-discuss/doru.serdin%40medias > at.ro > > > >
- Previous message (by thread): [members-discuss] GM topic
- Next message (by thread): [members-discuss] GM topic
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]